<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Rita Angelou on Medium]]></title>
        <description><![CDATA[Stories by Rita Angelou on Medium]]></description>
        <link>https://medium.com/@ritaaggelou?source=rss-79a3418e04a6------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*xjBKxBCleapWhffG</url>
            <title>Stories by Rita Angelou on Medium</title>
            <link>https://medium.com/@ritaaggelou?source=rss-79a3418e04a6------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 14:55:24 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ritaaggelou/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Power BI Tooltips: Add Context Without Cluttering Your Dashboard]]></title>
            <link>https://medium.com/microsoft-power-bi/power-bi-tooltips-add-context-without-cluttering-your-dashboard-31872491c0fc?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/31872491c0fc</guid>
            <category><![CDATA[dax]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[power-bi]]></category>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[power-bi-tutorials]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Fri, 24 Jul 2026 09:23:00 GMT</pubDate>
            <atom:updated>2026-07-28T10:55:05.558Z</atom:updated>
            <content:encoded><![CDATA[<p>Tooltip is a powerful feature in Power BI, yet it is often underused or misused. Instead of providing meaningful context, many tooltips simply repeat information that’s already visible in the report.</p><p>Imagine you’re presenting a sales dashboard comparing the performance of different shopping malls<strong>, when someone points at one of the bars and asks: “Is this better than last year?”</strong></p><p>The chart shows total sales, but it doesn’t tell the whole story. You need last year’s performance, growth percentage, and the categories driving that revenue.</p><p>Instead of opening another report page, we can answer all of those questions with a well-designed Power BI tooltip.</p><p>In this article, we’ll build a <strong>custom tooltip</strong> that transforms a simple bar chart into a compact performance dashboard, showing Year-to-Date sales, Year-over-Year growth, and the top-performing categories in just a few seconds.</p><p>Final Outcome:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O91_9HeRB50uzyLWiUULtA.jpeg" /></figure><p><strong>🎁 </strong><a href="https://powerbi-masterclass.short.gy/learning-database?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=database-medium-post-start"><strong>Get friend links for all of our 1500&gt; Power BI learning articles here</strong></a><strong> 🎁</strong></p><h3>A Sales Dashboard Example</h3><p>For this example I’m using the <strong>Shopping Mall Sales</strong> dataset from <a href="https://www.kaggle.com/datasets/mehmettahiraslan/customer-shopping-dataset">Kaggle</a>.</p><p>After cleaning the data, I organized it into a simple <strong>Star Schema</strong> consisting of:</p><ul><li>sales (fact table)</li><li>Date Table</li><li>shopping_mall</li><li>category</li><li>gender</li><li>payment_method</li></ul><p>If you’re unfamiliar with Star Schema modeling, I’ve linked another article at the end of this post.</p><p>This is my final model:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8Ltd_haO1xYQ_YAlpZSVQg.jpeg" /></figure><p>📌 <strong>Date Table:</strong></p><pre>DateTable = <br>ADDCOLUMNS (<br>    CALENDAR (DATE(2021,1,1), DATE(2023,12,31)),<br>    &quot;Year&quot;, YEAR([Date]),<br>    &quot;Month&quot;, MONTH([Date]),<br>    &quot;Month Name&quot;, FORMAT([Date], &quot;MMMM&quot;),<br>    &quot;Quarter&quot;, &quot;Q&quot; &amp; FORMAT([Date], &quot;Q&quot;)<br>)</pre><h3>Building the Main Visual</h3><p>First things first, let’s create the bar chart with total sales of each shopping mall for year 2022:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JkXZoezdd27ZbmdcCjNGIw.jpeg" /></figure><ol><li>Choose the <strong>stacked column chart</strong> visual.</li><li><strong>X-axis:</strong> shopping_mall (from shopping_mall table) and Year (from Date Table)</li><li>Filter year as <strong>2022</strong></li></ol><p>The goal of our tooltip is to show the growth or loss from last year and what people bought most from each shopping mall compared to last year.</p><p>📝 <strong>Note:</strong> In this example, the report is filtered to the full year of 2022, so the YTD value matches the annual total. I still use a YTD measure because it’s a best practice for time intelligence and will continue to work correctly if the report is later filtered to a specific month or date range.</p><h3>Building the Tooltip</h3><p>The goal of this tooltip is to provide users with a quick summary of each shopping mall’s performance when hovering over our main bar chart.</p><p>The tooltip will display:</p><ul><li><strong>Year-to-Date (YTD) Sales</strong></li><li><strong>Year-over-Year (YoY) Growth</strong></li><li><strong>Top 3 Best-Selling Categories</strong></li></ul><p>This allows you to compare this year’s performance with last year’s while also identifying the categories driving sales.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lcD3tSBE0CEpqUaQdCiZvw.jpeg" /></figure><h4>Step 1 : Create the Tooltip Page</h4><p>Let’s create a dedicated report page first that will be used as the tooltip.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EwRKX8MG1pIvMg64IbelTw.jpeg" /></figure><p>📌 <strong>Create a new page:</strong></p><ol><li>Click the <strong>+</strong> icon to create a new page.</li><li>Rename the page to <strong>Tooltip</strong>.</li></ol><p>📌 <strong>Enable Tooltip:</strong></p><ol><li>Open the <strong>Format</strong> pane.</li><li>Navigate to: <strong>Page Information → Allow use as tooltip → </strong>Turn the option <strong>On</strong>.</li></ol><p>This will convert your page into a tooltip page.</p><p>📝 Note: You can adjust the size from canvas setting if you do not like the current one.</p><h4>Step 2 : Create the Selected Mall Card</h4><p>Before showing any metrics, users should immediately know which shopping mall they’re looking at.</p><p>So, the first element that we are going to display in the tooltip is the shopping mall that we are hover over.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KPmotFFVYNov8MzeUTqGxw.jpeg" /></figure><p>📌 <strong>Create the measure:</strong></p><pre>Selected Mall = <br>SELECTEDVALUE(shopping_mall[shopping_mall])</pre><p>📌 <strong>Add the Card visual:</strong></p><ol><li>Insert a <strong>Card</strong> visual.</li></ol><p>2. Drag <strong>Selected Mall</strong> into the <strong>Value</strong> field.</p><p>3. Resize and position the card at the top of the tooltip.</p><h4>Step 3 : Create the Sales YTD Card (Show the Current Performance)</h4><p>The first KPI should answer the simplest question:</p><p><strong><em>“How much has this shopping mall sold this year?”</em></strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MGNjhJOeV59S70dvS1nCCQ.jpeg" /></figure><p>📌 <strong>Create the measure:</strong></p><pre>sales YTD = TOTALYTD(SUM(&#39;sales&#39;[sales]), &#39;DateTable&#39;[Date])</pre><p>📌 <strong>Add the Card visual:</strong></p><ol><li>Insert another <strong>Card</strong> visual.</li></ol><p>2. Add <strong>Sales YTD</strong> to the <strong>Value</strong> field.</p><p>3. Format the card as desired.</p><h4>Step 4 : Create the YoY Growth Card</h4><p>Just knowing total sales isn’t enough. We also want to know whether performance improved or declined compared to last year.</p><p>Therefore, a nice way to compare current Year-to-Date sales with the previous year’s performance is to create a YoY Growth KPI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_9XX1hfd7OSfzSZMJKlKMA.jpeg" /></figure><p>📌 <strong>Create the Previous Year Sales measure:</strong></p><pre>sales LY = CALCULATE (SUM(&#39;sales&#39;[sales]), SAMEPERIODLASTYEAR(&#39;DateTable&#39;[Date]))</pre><p>📌 <strong>Create the YoY Growth measure:</strong></p><pre>sales YoY KPI = DIVIDE ([sales YTD] - [sales LY], [sales LY], 0)</pre><p>📌 <strong>Add the KPI Card</strong></p><ol><li>Insert a <strong>Card</strong> visual.</li><li>Add <strong>Sales YoY KPI</strong> to the <strong>Value</strong> field.</li><li>Format the card as desired.</li></ol><p>📌 <strong>Apply Dynamic Colors</strong></p><p>To make positive growth easy to identify, we are going to create a measure that returns a color based on the YoY result.</p><p>📌 <strong>Create the color measure:</strong></p><pre>sales colour KPI = IF([sales YoY KPI]&gt;0, &quot;Green&quot;, &quot;Red&quot;)</pre><p>📌 <strong>Apply conditional formatting:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5zjYmkOwYrj_tvL5CAwPwQ.jpeg" /></figure><ol><li>Select the <strong>YoY Growth Card</strong>.</li><li>Navigate to: <strong>Format Visual → Callout Value → Color → Fx → Format style: Field value</strong> and set <strong>Based on field:</strong> <strong>sales colour KPI</strong></li></ol><p>Now, as you can check, positive values will appear in <strong>green</strong>, while negative values will appear in <strong>red</strong>.</p><p>📌 <strong>Add a Growth Indicator:</strong></p><p>We can further enhance the KPI by displaying an arrow indicating whether sales increased or decreased.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Sb2iBIzqHjQ4ru9sBaQ6RQ.jpeg" /></figure><p>📌 <strong>Create the indicator measure:</strong></p><pre>sales indicator PP KPI = var positive_icon = &quot;▲&quot; var negative_icon=&quot;▼&quot; var result = IF([sales YoY KPI]&gt;0, positive_icon, negative_icon) return result</pre><p>📌 <strong>Add the indicator:</strong></p><ol><li>Insert another <strong>Card</strong> visual.</li><li>Add <strong>Sales Indicator KPI</strong> to the <strong>Value</strong> field.</li><li>Resize and position the card.</li><li>Apply the same conditional formatting used for the YoY Growth card:</li></ol><p><strong>Format Visual → Callout Value → Color → Fx → Format style: Field value</strong> and set <strong>Based on field:</strong> <strong>sales colour KPI</strong></p><h4>Step 5 : Create the Top 3 Categories Matrix</h4><p>Finally, let’s answer the natural follow-up question:</p><p><strong>What’s driving those sales?</strong></p><p>To answer that question we should display the three product categories that generate the highest Year-to-Date sales for the selected shopping mall.</p><p>📌 <strong>Create the Category Rank measure:</strong></p><pre>Category Rank = <br>RANKX(<br>    ALL(category[category]),<br>    [sales YTD],<br>    ,<br>    DESC,<br>    DENSE<br>)</pre><p>This measure ranks categories based on their YTD sales.</p><p>📌 <strong>Create the Top 3 filter measure:</strong></p><pre>Show Top 3 Categories = <br>IF(<br>    [Category Rank] &lt;= 3,<br>    1<br>)</pre><p>This measure returns <strong>1</strong> for the top three ranked categories.</p><p>📌 <strong>Create the Top 3 sales by category:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Xtu0Vzcwzzbuo8U_A-u1LQ.jpeg" /></figure><ol><li>Insert <strong>Matrix</strong> visual.</li><li>Add <strong>category </strong>to the <strong>Rows </strong>field.</li><li>Add <strong>sales YTD, Sales indicator PP KPI </strong>and <strong>sales YoY KPI </strong>to the <strong>Values</strong> field.</li><li>Drag <strong>Show Top 3 Categories </strong>to the Filters area and format to show items when the value is <strong>1</strong></li><li>Resize and position the matrix.</li></ol><h3>Why This Still Works as a Tooltip?</h3><p>A common mistake many dashboard developers make is trying to fit an entire dashboard into a tooltip.</p><p>A good tooltip should answer <strong>one business question</strong>, not ten.</p><p>In our case, that question is:</p><p><strong><em>“How is this shopping mall performing?”</em></strong></p><p>Within a few seconds, users can answer:</p><p>✔️ How much has this mall sold?</p><p>✔️ Is it growing or declining?</p><p>✔️ Which categories generate the most revenue?</p><p>✔️ Which categories improved compared to last year?</p><p>Most importantly, users don’t need to leave the report or navigate to another page to find this information. They stay focused on the visual they’re exploring while getting the additional context they need.</p><h3>Final Thoughts</h3><p>Keep in mind that <strong>a tooltip shouldn’t try to replace a report page.</strong></p><p>Its purpose is to provide just enough context to answer the user’s next question without interrupting their workflow.</p><p>By combining a few KPIs, simple conditional formatting, and a compact matrix, we have transformed a standard Power BI <strong>tooltip</strong> into a miniature performance dashboard that <strong>users can understand in just a few seconds</strong>.</p><p>🌟 If you enjoyed this article don’t forget to 👏 clap and follow for more related content!</p><p>Related articles:</p><ul><li><a href="https://medium.com/microsoft-power-bi/tooltip-vs-drill-through-page-in-power-bi-a-practical-guide-to-better-report-design-e863b8114eaa">Tooltip vs Drill-Through Page in Power BI: A Practical Guide to Better Report Design</a></li><li><a href="https://medium.com/microsoft-power-bi/how-to-create-a-star-schema-in-power-bi-step-by-step-guide-7d6a895aa964">How to Create a Star Schema in Power BI (Step-by-Step Guide)</a></li><li><a href="https://medium.com/microsoft-power-bi/build-the-ultimate-calendar-table-in-power-bi-step-by-step-guide-0d09775ac6b4">Build the Ultimate Calendar Table in Power BI (Step-by-Step Guide)</a></li></ul><p><strong>💡 </strong><a href="https://powerbi-masterclass.short.gy/pbi-gpt?utm_source=medium&amp;utm_campaign=pbi-gpt-medium-post-end"><strong>MUST TRY — Power BI GPT — Personal Power BI Learning Coach</strong></a><strong>💡</strong></p><p><strong>Power BI Masterclass Article Classification</strong></p><p><strong>Level:</strong> Intermediate</p><p><strong>Category:</strong> Data Visualization, DAX</p><p><strong>Tags:</strong> Tutorial, Data Visualization, DAX</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=31872491c0fc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/microsoft-power-bi/power-bi-tooltips-add-context-without-cluttering-your-dashboard-31872491c0fc">Power BI Tooltips: Add Context Without Cluttering Your Dashboard</a> was originally published in <a href="https://medium.com/microsoft-power-bi">Power BI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Tooltip vs Drill-Through Page in Power BI: A Practical Guide to Better Report Design]]></title>
            <link>https://medium.com/microsoft-power-bi/tooltip-vs-drill-through-page-in-power-bi-a-practical-guide-to-better-report-design-e863b8114eaa?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/e863b8114eaa</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[power-bi]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[power-bi-guide]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Wed, 08 Jul 2026 07:43:19 GMT</pubDate>
            <atom:updated>2026-07-24T09:24:30.050Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yonZzh48Vx2Mg65wun2bcA.jpeg" /></figure><p>Most Power BI developers know how to create a tooltip or a drill-through page.</p><p>But, the real challenge isn’t building them, it is to know when to use each one.</p><p>I have seen many reports where users had to click though multiple pages to get a simple answer. And also others with so many tooltips that the dashboards feel cluttered.</p><p>In all of these cases the problem wasn’t technical but a design problem.</p><p>So, are you sure that you know the difference between Tooltip and Drill-Through and when to use each one?</p><p>In this article you’ll learn:</p><ul><li><strong>When to use a Tooltip and a Drill-Through</strong></li><li><strong>A practical example of each</strong></li><li><strong>Common mistakes that make reports difficult to use</strong></li></ul><h3>Understanding the Difference</h3><p><strong>Tooltips:</strong> Building a tooltip helps you to provide additional information without leaving the current page. That’s why its context should be quick and easy to understand.</p><p><strong>Drill-Through Page: A </strong>drill-through allow users to navigate to a dedicated page for a deeper analysis. It allows users to explore a topic in much greater detail.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6JJOYZ_50ZlsziNJAAE-0w.jpeg" /></figure><p>Can you spot the difference?</p><p>If not, think of it this way:</p><ul><li>Tooltip = Quick answer</li><li>Drill-Through = Detailed investigation</li></ul><p><strong>A tooltip should help the user to understand something immediately.</strong></p><p><strong>A drill-through page should help the user to explore something in depth.</strong></p><p>Let’s look at a simple example.</p><p><strong>🎁 </strong><a href="https://powerbi-masterclass.short.gy/learning-database?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=database-medium-post-start"><strong>Get friend links for all of our 1500&gt; Power BI learning articles here</strong></a><strong> 🎁</strong></p><h3>A Sales Dashboard Example</h3><p>The dataset that I am using for my example has this model view:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*k1xoAuBInrrTPQ_JFuan5g.jpeg" /></figure><p>To explain it quickly and easily: I have a dataset that contains sales information about different shopping malls.</p><p>Now, imagine you’re presenting a bar chart with total sales of each shopping mall for the current year during a management meeting.</p><p>One shopping mall immediately stands out because its sales are much lower than the others.</p><p>The first question comes:</p><p><strong>“How is this shopping mall performing?”</strong></p><p>At this point, you have two options:</p><ul><li>show additional information using a Tooltip</li><li>allow the user to Drill-Through to a dedicated analysis page</li></ul><p>That’s exactly where the choice between Tooltips and Drill-Through becomes important.</p><h3>Option 1: Create a Tooltip</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xjMgL5AaFqnzYLaREGhR_w.jpeg" /></figure><p>A great way to show the growth or loss from last year and maybe what people bought most from each shopping mall compared to last year would make for a simple and effective tooltip for this bar chart.</p><p><strong>My goal is to provide just enough information for the user to understand the mall’s performance without interrupting their workflow.</strong> So, when the user hovers over the bar chart would also have an idea about last year’s performance compared to this year and which categories were performing better.</p><p>So, my tooltip includes:</p><ul><li>Sales YTD</li><li>Year-over-Year Growth</li><li>Top 3 revenue-driving categories</li><li>Year-over-Year growth for each category</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ow7VTot0JsHq8K8lHCLEaA.jpeg" /></figure><p>But when is it time to change from a tooltip to a drill-through.</p><h3>Why it still works as a tooltip</h3><p>We said earlier that a tooltip shouldn’t contain too much information.</p><p>So, when does a tooltip become overloaded?</p><p>If after hovering over a shopping mall, I can answer these questions <strong>in just a few seconds</strong>:</p><ul><li>How much did this mall sell this year? ✔️</li><li>Is it growing or declining? ✔️</li><li>Which categories are driving revenue? ✔️</li><li>Which category is losing or gaining compared to last year? ✔️</li></ul><p>I don’t need to click anywhere or analyze multiple visuals. The tooltip gives me <strong>context</strong>, not <strong>analysis</strong>.</p><h3>When Does a Tooltip Become a Drill-Through Page?</h3><p>Although this tooltip contains multiple KPIs and a matrix, it still answers a single business question: <em>“How is this shopping mall performing?”</em></p><p>The user gets immediate context without leaving the report.</p><p>But eventually another question appears:</p><p><strong>“Why is it performing this way?”</strong></p><p>That’s where Drill-Through becomes the better choice.</p><h3>Option 2: Create a Drill-Through Page</h3><p>Unlike a tooltip, a drill-through page isn’t meant to provide a quick answer.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FC5n5N7c0nYYHzfLsgw7DA.jpeg" /></figure><p>With your drill-though you can provide a more detailed monthly overview, customer analysis, or transaction history for the current year.</p><p>In my example, right-clicking a shopping mall opens a dedicated report page containing:</p><ul><li>KPI summary</li><li>Monthly sales trend compared with last year</li><li>Category contribution</li><li>Average Basket by Gender</li><li>Transactions by age group</li><li>Sales by age group</li></ul><p>Rather than answering <strong>“How is this mall performing?”</strong>, the Drill-Through page helps us answer:</p><ul><li>When did performance start declining?</li><li>Which categories contributed the most?</li><li>Which categories performed better this year?</li><li>Which customer groups generated the highest sales?</li><li>Is the decline consistent throughout the year or concentrated in specific months?</li></ul><p>Notice how these questions require multiple visuals and comparisons.</p><p>Trying to fit all this information inside a tooltip would overwhelm the user.</p><p>By now, the difference between the two should be clear.</p><h3>Common Mistakes</h3><p>The most common mistakes to avoid when designing a dashboard are:</p><p><strong>❌ Using Drill-Through for simple questions.</strong></p><p>If the answer can fit inside a tooltip, don’t force users to navigate to another page.</p><p><strong>❌ Turning Tooltips into miniature reports.</strong></p><p>Tooltips should complement a visual, not replace an entire dashboard.</p><p><strong>❌ Repeating the same information.</strong></p><p>A tooltip and its Drill-Through page shouldn’t display identical content.</p><h3>Final Thoughts</h3><p>There isn’t a universal rule for deciding between Tooltips and Drill-Through pages.</p><p>The right choice depends on the question you’re trying to answer.</p><p>If users need a quick answer without leaving the report, use a Tooltip.</p><p>If they need to investigate trends, compare multiple dimensions, or understand the reasons behind a metric, use a Drill-Through page.</p><p><strong>Good Power BI reports aren’t defined by the number of features they use. They’re defined by how naturally users can find the answers they’re looking for.</strong></p><p>In my next few articles I will focus on how to build the above visuals (Tooltip and Drill-Through Page).</p><p>So, if you liked this article and want to learn more please don’t forget to follow!</p><p>Related articles:</p><p><a href="https://medium.com/@ritaaggelou/power-bi-tooltips-add-context-without-cluttering-your-dashboard-31872491c0fc">Power BI Tooltips: Add Context Without Cluttering Your Dashboard</a></p><p><strong>💡 </strong><a href="https://powerbi-masterclass.short.gy/pbi-gpt?utm_source=medium&amp;utm_campaign=pbi-gpt-medium-post-end"><strong>MUST TRY — Power BI GPT — Personal Power BI Learning Coach</strong></a><strong>💡</strong></p><p><strong>Power BI Masterclass Article Classification</strong></p><p><strong>Level:</strong> Beginner</p><p><strong>Category:</strong> Data Visualization</p><p><strong>Tags:</strong> Guide, Data Visualization</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e863b8114eaa" width="1" height="1" alt=""><hr><p><a href="https://medium.com/microsoft-power-bi/tooltip-vs-drill-through-page-in-power-bi-a-practical-guide-to-better-report-design-e863b8114eaa">Tooltip vs Drill-Through Page in Power BI: A Practical Guide to Better Report Design</a> was originally published in <a href="https://medium.com/microsoft-power-bi">Power BI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ribbon Chart in Power BI: Track Category Rankings Over Time]]></title>
            <link>https://medium.com/microsoft-power-bi/ribbon-chart-in-power-bi-track-category-rankings-over-time-693f299579f4?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/693f299579f4</guid>
            <category><![CDATA[power-bi-tutorials]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[power-bi]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Tue, 09 Jun 2026 06:43:35 GMT</pubDate>
            <atom:updated>2026-06-09T20:02:53.965Z</atom:updated>
            <content:encoded><![CDATA[<p>When I want to compare category performance over time, my first instinct is usually to create a line chart or a stacked column chart. They work well, and most of us use them regularly.</p><p>But sometimes the question isn’t just <em>how much did sales change?</em></p><p>The real question is <em>who was winning each month?</em></p><p>That’s where Ribbon Charts become interesting. They are one of those Power BI visuals that many people overlook, but they can tell a story that would be difficult to spot in a traditional chart.</p><p><strong>A Ribbon Chart</strong> is similar to a stacked column chart, but instead of keeping categories in a fixed legend order, it <strong>dynamically reorders them based on their value at each point in time.</strong> As a result, the visual clearly shows which category is leading and how rankings change throughout the period.</p><p>Think of it as watching a race unfold. As competitors overtake one another, their positions change, making it easy to identify leaders, challengers, and periods of dominance.</p><p>Final Dashboard:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pux60nr1OigbhvDZFXJHJw.jpeg" /></figure><h3>Dataset Used</h3><p>For this example, I used a restaurant sales dataset from Kaggle. The only columns required are:</p><ul><li>Restaurant Type (Cafe, Casual Dining, Fine Dining, Food Stall, and Kopitiam)</li><li>Sales</li><li>Date</li></ul><p>Before proceeding, ensure these fields do not contain missing values. Null values can affect rankings and produce misleading results in the ribbon chart.</p><h3>Dashboard Objective</h3><p>When I explored the dataset, I wasn’t particularly interested in total yearly sales. What caught my attention was how the leadership changed throughout the year.</p><p>Did one restaurant type dominate from start to finish? Or did different categories take turns being the top performer?</p><p>That became the main question behind this dashboard:</p><p><strong>Which restaurant type generated the highest sales each month, and how did leadership change throughout the year?</strong></p><p>To answer this, we’ll build a dashboard that includes:</p><ul><li>A Ribbon Chart showing monthly ranking changes</li><li>A Top Insight card</li><li>A Most Volatile Category card</li><li>A Highest Annual Sales card</li><li>A summary table showing Months ranked #1 and Total annual sales</li></ul><p>Together, these visuals provide both the big picture and supporting details.</p><p>This will be our final dashboard:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vuuyZ0mAvx6cDth9M0Ve2w.jpeg" /><figcaption>Final Dashboard Overview</figcaption></figure><p><strong>🎁 </strong><a href="https://powerbi-masterclass.short.gy/learning-database?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=database-medium-post-start"><strong>Get friend links for all of our 1500&gt; Power BI learning articles here</strong></a><strong> 🎁</strong></p><h3>Create Date Table</h3><p>I always create a Date Table before building any time-based visual. In this example it might feel unnecessary because we only have one year of data, but it gives us much more flexibility when creating rankings and time calculations later.</p><p>The Date Table:</p><pre>DateTable = <br>ADDCOLUMNS(<br>    CALENDAR(DATE(2024,1,1), DATE(2024,12,31)),<br>    &quot;Year&quot;, YEAR([Date]),<br>    &quot;Month&quot;, MONTH([Date]),<br>    &quot;Month Name&quot;, FORMAT([Date], &quot;MMMM&quot;),<br>    &quot;Quarter&quot;, &quot;Q&quot; &amp; FORMAT([Date], &quot;Q&quot;)<br>)</pre><p>After creating the table:</p><ol><li>Create a relationship between the dataset Date field and DateTable[Date].</li><li>Sort Month Name by Month Number to ensure months appear chronologically.</li></ol><h3>Create a Measures Table</h3><p>One thing I do in almost every Power BI project is create a dedicated Measures table.</p><p>Is it absolutely necessary? No.</p><p>But once a report starts growing, having all your measures in one place makes life much easier.</p><p>To do this create a new table:</p><pre>Measures_Table = {BLANK()}</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*suwFq2bYgqZk_9KHXRLcVg.jpeg" /><figcaption>Final Data Model</figcaption></figure><h3>Build the Ribbon Chart</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LMT9auB0CT8q1wI3MR30Ew.jpeg" /><figcaption>Ribbon Chart Configuration</figcaption></figure><h4>Configure the Visual</h4><p>It is rather easy to create the ribbon chart, follow the steps below:</p><ol><li>Choose the Ribbon chart icon from the visualizations pane.</li><li>Add the columns:</li></ol><ul><li>X-axis: Month Name (from date table)</li><li>Y-axis: Sales (sum of sales from your dataset)</li><li>Legend: Restaurant Type (from your dataset)</li></ul><p>This is usually the moment when the Ribbon Chart starts to make sense.</p><p>With a line chart, you would have to inspect the values carefully to identify leadership changes. Here, your eyes immediately follow the category that sits on top and notice when another category overtakes it.</p><h4>Optional: Filter Categories That Never Led Sales</h4><p>In my dataset there were five restaurant types, but two of them never managed to reach the top position during the year.</p><p>I could have left them in the visual, but I felt they added unnecessary clutter. Since the goal of this dashboard is to focus on leadership changes, I decided to keep only the categories that actually competed for first place.</p><ol><li>First create a ranking measure:</li></ol><pre>Restaurant Rank = <br>RANKX(<br>    ALL(restaurant_sales[restaurant_type]),<br>    CALCULATE(SUM(restaurant_sales[Sales])) +<br>        DIVIDE(<br>            RANKX(<br>                ALL(restaurant_sales[restaurant_type]),<br>                MAX(restaurant_sales[restaurant_type]),<br>                ,<br>                ASC<br>            ),<br>            1000000<br>        ),<br>    ,<br>    DESC,<br>    DENSE<br>)</pre><p>2. Then calculate the number of months each category ranked first:</p><pre>Months At Rank 1 = <br>SUMX(<br>    VALUES(DateTable[Month Name]),<br>    IF([Restaurant Rank] = 1, 1, 0)<br>)</pre><p>3. Finally, create a filter measure:</p><pre>Is Leader = <br>VAR Wins =<br>    CALCULATE(<br>        [Months At Rank 1],<br>        ALLEXCEPT(<br>            restaurant_sales,<br>            restaurant_sales[restaurant_type]<br>        )<br>    )<br>RETURN<br>IF(Wins &gt; 0, &quot;Yes&quot;, &quot;No&quot;)</pre><p>Apply this measure as a visual filter and keep only values containing “Yes”.</p><h4>Enable Data Labels</h4><p>To make rankings easier to read:</p><ul><li>Select the Ribbon Chart</li><li>Open the Format pane</li><li>Turn Data Labels On</li></ul><h3>Add KPI Cards</h3><p>At this point, the Ribbon Chart already answers our main question.</p><p>However, I like adding a few supporting KPIs because they help readers understand the story faster without having to inspect the chart in detail.</p><p>I chose to create 3 card visuals (Top Insight, Most Volatile Category and Highest Annual Sales) and a table that will show all the restaurant types that are at least one month in top with how many times the were on top and their total sales of the year.</p><p>That way our dashboard answers more clearly and quickly the question that we made in first place.</p><h4>Top Insight</h4><p>This card highlights the category that spent the most months in first place.</p><p>Create the measure:</p><pre>Top Insight = <br>VAR TopRestaurant =<br>    TOPN(<br>        1,<br>        VALUES(restaurant_sales[restaurant_type]),<br>        [Months At Rank 1],<br>        DESC<br>    )<br><br>VAR RestaurantName =<br>    MAXX(TopRestaurant, restaurant_sales[restaurant_type])<br><br>VAR TotalMonths =<br>    MAXX(TopRestaurant, [Months At Rank 1])<br><br>RETURN<br>RestaurantName &amp;<br>&quot; ranked #1 in &quot; &amp;<br>TotalMonths &amp;<br>&quot; months&quot;</pre><p>Choose Card from Visualizations Pane → Value: Top Insight</p><h4>Most Volatile Category</h4><p>Calculate categories that experienced the most ranking changes throughout the year.</p><p>Create the measures:</p><pre>Rank Changes = <br>VAR CurrentRank = [Restaurant Rank]<br><br>VAR PreviousRank =<br>    CALCULATE(<br>        [Restaurant Rank],<br>        DATEADD(DateTable[Date], -1, MONTH)<br>    )<br><br>RETURN<br>IF(<br>    NOT ISBLANK(PreviousRank) &amp;&amp;<br>    CurrentRank &lt;&gt; PreviousRank,<br>    1,<br>    0<br>)</pre><pre>Total Rank Changes = <br>SUMX(<br>    VALUES(DateTable[Month Name]),<br>    [Rank Changes]<br>)</pre><pre>Most Volatile Category = <br>VAR TopRestaurant =<br>    TOPN(<br>        1,<br>        VALUES(restaurant_sales[restaurant_type]),<br>        [Total Rank Changes],<br>        DESC<br>    )<br><br>RETURN<br>MAXX(<br>    TopRestaurant,<br>    restaurant_sales[restaurant_type]<br>)</pre><p>Choose Card from Visualizations Pane → Value: Most Volatile Category</p><h4>Highest Annual Sales</h4><pre>Highest Annual Sales = <br>VAR TopRestaurant =<br>    TOPN(<br>        1,<br>        VALUES(restaurant_sales[restaurant_type]),<br>        [Total Sales],<br>        DESC<br>    )<br><br>VAR RestaurantName =<br>    MAXX(<br>        TopRestaurant,<br>        restaurant_sales[restaurant_type]<br>    )<br><br>VAR SalesAmount =<br>    MAXX(<br>        TopRestaurant,<br>        [Total Sales]<br>    )<br><br>RETURN<br>RestaurantName &amp;<br>&quot; - &quot; &amp;<br>FORMAT(SalesAmount, &quot;$#,##0,,.0M&quot;)</pre><p>Choose Card from Visualizations Pane → Value: Highest Annual Sales</p><h3>Create the Summary Table</h3><p>Finally, add a Table visual containing:</p><ul><li>Restaurant Type</li><li>Months At Rank 1</li><li>Total Sales</li></ul><p>This table provides a detailed breakdown of each category’s performance and complements the high-level insights shown in the KPI cards.</p><h3>Final Thoughts</h3><p>Ribbon Charts are not the most popular visual in Power BI, and honestly, there are many situations where a simple line chart is the better choice.</p><p>However, when your goal is to track ranking changes over time, Ribbon Charts can be incredibly effective. They make it easy to see who is leading, when leadership changes, and which categories consistently stay near the top.</p><p>For this dashboard, a line chart would have shown the sales trends. A Ribbon Chart tells the competitive story behind those trends and that’s exactly why I chose it.</p><p>If you found this article helpful, feel free to follow for more Power BI and data visualization tutorials.</p><p><strong>💡 </strong><a href="https://powerbi-masterclass.short.gy/pbi-gpt?utm_source=medium&amp;utm_campaign=pbi-gpt-medium-post-end"><strong>MUST TRY — Power BI GPT — Personal Power BI Learning Coach</strong></a><strong>💡</strong></p><p><strong>Power BI Masterclass Article Classification</strong></p><p><strong>Level:</strong> Beginner</p><p><strong>Category:</strong> Data Visualization</p><p><strong>Tags:</strong> Tutorial, Data Visualization</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=693f299579f4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/microsoft-power-bi/ribbon-chart-in-power-bi-track-category-rankings-over-time-693f299579f4">Ribbon Chart in Power BI: Track Category Rankings Over Time</a> was originally published in <a href="https://medium.com/microsoft-power-bi">Power BI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Same Data, 4 Different Stories: Understanding Power BI Column Charts]]></title>
            <link>https://medium.com/learning-data/the-same-data-4-different-stories-understanding-power-bi-column-charts-7481a99c8829?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/7481a99c8829</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[business-intelligence]]></category>
            <category><![CDATA[power-bi]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[dashboard-design]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Fri, 22 May 2026 07:09:47 GMT</pubDate>
            <atom:updated>2026-06-18T21:41:03.949Z</atom:updated>
            <content:encoded><![CDATA[<p>Column charts are some of the most commonly used visuals in Power BI dashboards and reports.</p><p>They are simple and extremely effective when you want to compare values across categories.</p><p>However, many Power BI users struggle with one question:</p><blockquote><strong><em>Which type of column chart should I use?</em></strong></blockquote><p>There are 4 main types:</p><ul><li><strong>Single Column Chart</strong></li><li><strong>Clustered Column Chart</strong></li><li><strong>Stacked Column Chart</strong></li><li><strong>100% Stacked Column Chart</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GILAIKejt8DxuHA84krq3g.jpeg" /></figure><p>At first glance, they may look very similar but each chart tells a completely different story, even when using the same data.</p><p>I’ve personally found myself confused many times about which one to choose for the right storytelling. But once you understand what each chart emphasizes, choosing the right one becomes much easier.</p><p>To explain this, I’ll use a fitness activity dataset containing:</p><ul><li>Very Active Minutes</li><li>Fairly Active Minutes</li><li>Lightly Active Minutes</li><li>Daily activity records filtered by month and year</li></ul><p>Using the same dataset, we’ll see how each chart answers different questions and when each one should be used in Power BI dashboards.</p><p>In this article, we are going to clarify:</p><ul><li>When to use each chart type.</li><li>What questions it answers well.</li><li>What limitations it has.</li><li>And when you should use it in Power BI dashboards.</li></ul><h3>1. Simple Column Chart</h3><p>Simple Column Chart is the chart that you can use to compare numerical values across different categories.</p><p>For our example, we will use it to visualize the overall activity volume for the first week of our month.</p><p>This chart will give us a quick overview of our overall activity and immediately will help us answer several important questions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jmOQ-EgBu-R1vMajVY4WdA.jpeg" /></figure><p>By looking at the chart, we can instantly answer questions like:</p><ul><li>Which day was the most active?</li><li>Which day was the least active?</li><li>Are activity levels increasing during the week?</li><li>Are weekends more active than weekdays?</li></ul><p>For example, in our dashboard, Thursday clearly had the highest total activity.</p><p>This type of column chart works great for:</p><ul><li>Comparing total activity</li><li>Spotting high and low activity</li><li>Tracking trends over time</li></ul><p>As you can notice by now, it is an easy way to read and focuses only on one thing: <strong>overall performance</strong>.</p><p>Now let’s say that we want to do a deeper analysis.</p><p>What happens if we want to check not only the total activity but to see how much very active time was recorded or to check if the activity was mostly light or intense throughout the day.</p><p>To answer those questions, we need to break the data into categories (Very Active, Fairly Active and Lightly Active)</p><p>That’s where the Clustered Column chart becomes useful.</p><h3>2. Clustered Column Chart</h3><p>A Clustered Column chart helps you display data within categories side-by-side so you can easily compare groups.</p><p>So, instead of just showing the total activity we can now separate it into categories:</p><ul><li>Very Active Minutes</li><li>Fairly Active Minutes</li><li>Lightly Active Minutes</li></ul><p>This will help us to have a closer look of our activity per day.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AcKCmfveLTuuJvtCWeftCQ.jpeg" /></figure><p>Now we can clearly see that most of our activity comes from Lightly Active Minutes, while Sunday and Friday contain more Very Active Minutes than the rest of the week.</p><p>This chart helps answer questions like:</p><ul><li>Which activity type is highest each day?</li><li>On which day was Very Active time highest?</li><li>How do Fairly Active Minutes compare across days?</li><li>Which category dominates the week?</li></ul><p>This type of chart works great for:</p><ul><li>Comparing exact values.</li><li>Identifying dominant categories.</li><li>Analyzing differences between activity types.</li></ul><p>However, even though category comparison becomes very easy, another problem appears.</p><p>Try comparing the total activity between Sunday, Monday, and Tuesday.</p><p>It’s not that clear, right?</p><p>Because the categories are separated into multiple bars, our brain has to mentally add them together to estimate the total.</p><p>That is the main limitation of the Clustered Column Chart.</p><p>It is excellent for comparing categories, but not so great for comparing totals.</p><p>So if your goal is to compare categories while also keeping total activity easy to understand, the next chart type is a much better choice.</p><h3>3. Stacked Column Chart</h3><p>A Stacked Column chart combines all categories into a single bar.</p><p>Instead of placing categories side by side, it stacks them on top of each other.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AIOIubWGD_9ja1AEDVbefg.jpeg" /></figure><p>With this chart, it is easier to compare the total activity while you can see the contribution of each category.</p><p>In other words, it combines overview and detail in the same visual.</p><p>Now try to identify:</p><ul><li>Which day had the highest total activity?</li><li>Which category contributed most to the total?</li><li>Whether activity was mostly light or intense?</li></ul><p>The answer comes naturally.</p><p>This type of chart works great for:</p><ul><li>Showing totals and composition together.</li><li>Understanding category contribution.</li><li>Comparing overall performance while keeping breakdowns visible.</li></ul><p>A Stacked Column chart is better than a Simple Column chart when you want to provide more context without losing the bigger picture.</p><p>However, it has limitations.</p><p>For example, try to compare Fairly Active Minutes between Monday and Saturday.</p><p>It’s not that clear.</p><p>That’s because only the bottom segment shares the same baseline.</p><p>Still, for many dashboards, this chart provides one of the best balances between detail and readability.</p><p>But what if totals are no longer important?</p><p>What if we only care about percentages and proportions?</p><p>That’s where the 100% Stacked Column Chart becomes extremely useful.</p><h3>4. 100% Stacked Column Chart</h3><p>A 100% Stacked Column Chart normalizes every bar to the same height.</p><p>Instead of showing actual totals, it shows the percentage contribution of each category.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fBtV8w9ui8zvDkvMrW8Clw.jpeg" /></figure><p>Now the focus is no longer:</p><ul><li><em>Which day had the most activity?</em></li></ul><p>Instead, the focus becomes:</p><ul><li><em>How was the activity distributed?</em></li></ul><p>This chart is perfect when you want to understand composition and proportions.</p><p>This chart will help you answer questions like:</p><ul><li>Which day had the highest proportion of intense activity?</li><li>Which days were mostly lightly active?</li><li>How does activity composition change during the week?</li><li>Which day had the most balanced activity distribution?</li></ul><p>This type of chart works great for:</p><ul><li>Comparing percentages</li><li>Understanding relative contribution</li><li>Analyzing composition patterns</li></ul><p>However, there is one major tradeoff:</p><p>You completely lose visibility of totals.</p><p>A day with 200 active minutes and a day with 80 active minutes can look identical if their proportions are the same.</p><p>That’s why this chart should only be used when percentages matter more than actual values.</p><h3>Final Thoughts</h3><p>One of the most important lessons in dashboard design is this:</p><blockquote><strong><em>The same data can tell completely different stories depending on the chart you choose.</em></strong></blockquote><p>So, by choosing the right visualization you not only make a design decision you make a communication decision.</p><p>Therefore, before selecting a chart in your Power BI dashboard, ask yourself:</p><ul><li>Do I care about totals?</li><li>Do I care about exact comparisons?</li><li>Do I care about composition?</li><li>Or do I care about percentages?</li></ul><p>The answer will usually tell you exactly which chart to use.</p><p>Once you understand that, creating more effective and insightful dashboards becomes much easier.</p><p>If you found this article helpful, feel free to follow for more Power BI and data visualization tutorials.</p><p><em>The contents of external submissions are not necessarily reflective of the opinions or work of </em><a href="http://mavenanalytics.io/"><em>Maven Analytics</em></a><em> or any of its team members.</em></p><p><em>We believe in fostering lifelong learning and our intent is to provide a platform for the data community to share their work and seek feedback from the Maven Analytics data fam.</em></p><p><a href="https://medium.com/learning-data/how-to-get-your-work-published-by-learning-data-with-maven-analytics-7df21e466a3e?source=friends_link&amp;sk=020dfac485597d602e218968d9ffb395"><em>Submit your own writing here</em></a><em> if you’d like to become a contributor.</em></p><p><em>Happy learning!</em></p><p><em>-Team Maven</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7481a99c8829" width="1" height="1" alt=""><hr><p><a href="https://medium.com/learning-data/the-same-data-4-different-stories-understanding-power-bi-column-charts-7481a99c8829">The Same Data, 4 Different Stories: Understanding Power BI Column Charts</a> was originally published in <a href="https://medium.com/learning-data">Learning Data</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Create a Radial Chart in Power BI (Step-by-Step Guide)]]></title>
            <link>https://medium.com/microsoft-power-bi/how-to-create-a-radial-chart-in-power-bi-step-by-step-guide-390fe8da920b?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/390fe8da920b</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[power-bi-tutorials]]></category>
            <category><![CDATA[power-bi]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Tue, 17 Mar 2026 07:32:32 GMT</pubDate>
            <atom:updated>2026-03-18T13:54:04.423Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*y4c6A_oGZx-mP6n6tC_b6A.png" /><figcaption>Image created by Author</figcaption></figure><p>Most Power BI dashboards rely on the same visuals such as bar charts, line charts and tables.While these work well, they don’t always capture the attention your data deserves.</p><p>So, what if you could present your data in a more engaging way?</p><p>In this article, we’ll explore how to create a <strong>radial chart in Power BI, </strong>a<strong> </strong>circular visualization that transforms simple revenue data into a compelling story. Using real-world data from the <em>Fortune Global 500</em>, you’ll learn how to build this chart step by step, even though Power BI does not offer it as a native visual.</p><p><strong>🎁 </strong><a href="https://powerbi-masterclass.short.gy/learning-database?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=database-medium-post-start"><strong>Get friend links for all of our 1500&gt; Power BI learning articles here</strong></a><strong> 🎁</strong></p><h3>Creating the Dataset</h3><p>Open a blank Power BI report and select <strong>Get Data → Web</strong>. Then copy and paste the following URL:</p><p><a href="https://en.wikipedia.org/wiki/Fortune_Global_500">Fortune Global 500 - Wikipedia</a></p><p>The Fortune Global 500, also known as the Global 500<em>,</em> is an annual ranking of the top 500 corporations worldwide based on total revenue.</p><p>Select the necessary tables and click <strong>Load</strong>.<br> In this example, we will load the following tables:</p><ul><li>Fortune Global 500 list of <strong>2020</strong></li><li>Fortune Global 500 list of <strong>2021</strong></li><li>Fortune Global 500 list of <strong>2022</strong></li><li>Fortune Global 500 list of <strong>2023</strong></li><li>Fortune Global 500 list of <strong>2024</strong></li></ul><p>Each table contains the <strong>top 10 companies ranked by revenue (in billions of USD)</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0jpX1s56Rv-ydWAnEfPx9Q.jpeg" /><figcaption>Image created by Author</figcaption></figure><h3>Appending the Tables and Cleaning the Fact Table</h3><h4>Appending the Tables</h4><p>To simplify the data model, I wanted to store all the data in <strong>one fact table</strong>. Appending the tables allows us to store all yearly data in a single table, which makes filtering by year much easier in Power BI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gO0bsoUT6Eh51TIQkQpCOA.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>Step 1 : Add a Year Column</strong></p><p>Select the table Fortune Global 500 list of <strong>2024</strong> and:</p><ol><li>Go to <strong>Add Column</strong> → Choose <strong>Custom Column</strong></li><li>Enter the following in the window that opens. <br>New column name: <strong>Year </strong>and Custom column formula: <strong>= 2024</strong></li></ol><p>Repeat this process for each table using the corresponding year.</p><p><strong>Step 2 : Append the Tables</strong></p><p>Since the structure of all tables is identical, we can combine them using <strong>Append Queries</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bTP3oD1_YOEmfIqB9uFGHQ.jpeg" /><figcaption>Image created by Author</figcaption></figure><ol><li>In Power Query, go to <strong>Home → Append Queries → Append Queries as New</strong>.</li><li>Select all the tables and click OK.</li></ol><p>This creates a new table that contains all records from the individual tables.<br> I renamed this table: <strong>Fact_FortuneRevenue</strong></p><h4>Cleaning the Fact Table</h4><p>Next, we need to clean the table before we use it in Power BI visuals.</p><p><strong>Step 1 : Filter the Rank Column</strong></p><p>Since the radial chart becomes difficult to read with too many categories, we limit the visualization to the<strong> top five companies per year</strong>.</p><p>In the <strong>Rank</strong> column, keep only values <strong>1 to 5</strong> so that we focus on the top five companies for each year.</p><p><strong>Step 2 : Convert Revenue to a Numeric Column</strong></p><p>The <strong>Revenue in USD</strong> column is currently stored as text and contains symbols such as<em> $</em> and the word <em>billion</em>. These need to be removed before converting the column to a numeric data type.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Bhqz4DDIrsMCUBdW3PGwwA.png" /><figcaption>Image created by Author</figcaption></figure><ol><li>Right-Click the column and Choose Replace Values:</li></ol><ul><li>Value to Find : <strong>$ </strong>Replace With: <strong>(Blank)</strong></li><li>Value to Find : <strong>billion </strong>Replace With: <strong>(Blank)</strong></li><li>Replace Values: <strong>.</strong> <strong>(dot)</strong> Replace With:<strong> , (comma)</strong></li></ul><p>2. Change Type: <strong>Decimal Number</strong>.</p><p><strong>Step 3 : Fix Text Inconsistencies in the Industry Column</strong></p><p>When inspecting the <strong>Industry</strong> column, we can observe inconsistent text formatting. For example: <br>“Internet Services and Retailing ” and “Internet services and retailing”.</p><p>Although these values represent the <strong>same industry</strong>, Power BI treats them as <strong>different categories</strong>, which can create problems in the data model.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kp4oQ5NAScKEnVyTXQ0RIA.png" /><figcaption>Image created by Author</figcaption></figure><p>To standardize the formatting:</p><ol><li>Select<strong> Industry Column → Go to Transform</strong></li><li>Choose <strong>Format → Capitalize Each Word</strong></li></ol><p>This ensures that all values follow a consistent format.</p><p>Finally, we are done with the cleaning process, so click <strong>Close &amp; Apply</strong>.</p><h4>Hiding Unnecessary Tables</h4><p>The original tables used during the data preparation stage are no longer needed in the report view.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lFhtHdVTWVZqH2CploKw6w.jpeg" /><figcaption>Image created by Author</figcaption></figure><p>To keep the model clean and not get confused:</p><ol><li>Go to <strong>Table View</strong></li><li>Right-click the table → Select <strong>Hide in report view </strong>or<strong> Delete from model</strong></li></ol><p>Repeat this step for the other tables.</p><h3>Building your Model</h3><p>💡 You can also create a star schema for your model for practise. But as this dataset is pretty small you will not need to. This is my approach to this example:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5prPstYMOccR6bSV0ZROsA.png" /><figcaption>Image created by Author</figcaption></figure><p>For this example, I will continue using only the fact table.</p><h3>Creating the Measures for the Radial Chart</h3><p>Since Power BI doesn’t have a native radial chart visual, we will simulate the effect using a custom visual structure.</p><p>To build the Radial Chart we need to create some DAX measures that help us with the shape of the visualization.</p><h4>1. Total Revenue</h4><p>First, we are going to create a measure that will calculate total revenue.</p><pre>Total Revenue = SUM(Fact_FortuneRevenue[Revenue in USD])</pre><p>This measure aggregates the revenue values from the fact table.</p><h4>2. Top Revenue</h4><p>Next, we want to identify the company with the highest revenue for each year.</p><pre>Top Revenue = <br>CALCULATE(<br>    MAX(Fact_FortuneRevenue[Revenue in USD]),<br>    Fact_FortuneRevenue[Rank] = 1,<br>    ALLEXCEPT(Fact_FortuneRevenue, Fact_FortuneRevenue[Year])<br>)</pre><p>This measure finds the<strong> maximum revenue value</strong> <strong>for the company</strong> <strong>ranked number 1 </strong>in the selected year.</p><p>Since our dataset includes a <strong>Year </strong>column, the result automatically adjusts when a different year is selected in the report.</p><h4>3. Target Measure</h4><p>The target measure represent the difference between the <strong>total revenue of the top five companies</strong> and the <strong>revenue of the top-ranked company</strong>.</p><pre>Target = <br>CALCULATE(<br>    [Total Revenue],<br>    ALLEXCEPT(Fact_FortuneRevenue, Fact_FortuneRevenue[Year])<br>) - [Top Revenue]</pre><p>This value helps define the background structure used to create the radial layout.</p><h4>4. Hidden Measure</h4><p>Now, we will create a measure that will help us create the necessary <strong>spacing in the radial chart</strong>.</p><pre>Hidden = ([Target]+[Top Revenue])*0.33</pre><p>Although this measure does not represent meaningful data, it plays an important role in shaping the final visual.</p><h4>5. Title of the Chart</h4><p>Finally, the last measure will help us create a dynamic title for our chart.</p><pre>Title = &quot;Top 5 Companies by Revenue ($B) - &quot; &amp; SELECTEDVALUE(Fact_FortuneRevenue[Year])</pre><p>This measure displays the <strong>year currently selected in the report</strong>.</p><h3>Creating the Background</h3><p>In the image below, I show how I created the background layout to ensure that the chart appears clean and that the measures work correctly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*H6usWjsnY1Ck5Ah_p5M5Lg.png" /><figcaption>Image created by Author</figcaption></figure><p>Now that we have prepared our dataset and created the required measures and background, we can move on to building the radial chart in Power BI.</p><h3>Creating the Radial Chart</h3><p>To simulate a radial chart in Power BI, we will use <strong>multiple donut charts layered on top of each other</strong>.</p><h4>Step 1: Create the First Donut Chart (Rank 1)</h4><p>This donut chart represents <strong>Rank 1 (the company with the highest revenue).</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BW-AXiqFzZ-d-TdhivwheQ.png" /><figcaption>Image created by Author</figcaption></figure><ol><li>Insert a <strong>Donut Chart</strong> and</li><li>Add the following in the <strong>Values </strong>area:</li></ol><ul><li><strong>Revenue in USD (MAX)</strong></li><li><strong>Target</strong></li><li><strong>Hidden</strong></li></ul><p>3. Apply the following filter:<br> <strong>Rank → Basic Filtering → Select 1</strong></p><h4>Step 2: Format the Chart</h4><p>➡️ <strong>Visual Settings</strong></p><ol><li>Turn the following off:</li></ol><ul><li>Legend</li><li>Detail labels</li></ul><p>2. Set the color:</p><ul><li>Slices → Color → Max of Revenue in USD: #8F1238, Target : #999999 and Hidden : #808080</li></ul><p>I selected these colors to create a <strong>clean and professional look</strong>.</p><p>💡 If you choose different colors, make sure the <strong>Hidden measure uses the same color as the background shape</strong>, so that it becomes invisible.</p><p>3. Set Spacing:</p><p>Spacing → Inner radius : 90%</p><p>➡️ <strong>General Settings</strong></p><p>Disable the following:</p><ul><li>Title</li><li>Header Icons</li><li>Tooltips</li><li>Background (Effects)</li></ul><h4>Creating the Remaining Rings</h4><p>To create the radial effect, <strong>duplicate the first donut chart</strong> and adjust the filter, color, and inner radius for each rank.</p><p>➡️ <strong>Rank 2</strong></p><p>Copy and paste the first chart and apply the following changes:</p><p><strong>Filter:</strong></p><p>Rank → Basic Filtering → Select 2</p><p><strong>Slice Color:</strong></p><p>Max of Revenue in USD: #CC3A63 <br>(Target and Hidden remain the same color)</p><p><strong>Spacing:</strong></p><p>Inner radius : 88%</p><p>Resize the chart so it fits <strong>inside the first donut chart</strong>.</p><p>➡️ <strong>Rank 3</strong></p><p><strong>Filter:</strong></p><p>Rank → Basic Filtering → Select 3</p><p><strong>Slice Color:</strong></p><p>Max of Revenue in USD: #E6678C</p><p><strong>Spacing:</strong></p><p>Inner radius : 84%</p><p>➡️ <strong>Rank 4</strong></p><p><strong>Filter:</strong></p><p>Rank → Basic Filtering → Select 4</p><p><strong>Slice Color:</strong></p><p>Max of Revenue in USD: #F29AB5</p><p><strong>Spacing:</strong></p><p>Inner radius : 81%</p><p>➡️ <strong>Rank 5</strong></p><p><strong>Filter:</strong></p><p>Rank → Basic Filtering → Select 5</p><p><strong>Slice Color:</strong></p><p>Max of Revenue in USD: #F8C8D8</p><p><strong>Spacing:</strong></p><p>Inner radius : 73%</p><p>Once all donut charts are layered and aligned, they will create the <strong>final radial chart visualization</strong>.</p><p>You should have a visual that would look like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nOPuXm0xlnhgEkYCKACs-A.png" /><figcaption>Image created by Author</figcaption></figure><h3>Creating the Multi-Row Cards</h3><p>At this point, the radial chart is complete, but it does not display detailed information. Therefore, we are going to build multi-row cards to improve this.</p><h4>Step 1: Create the First Card (Rank 1)</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BK7vC8iitFmOIpsVdSVB_w.png" /><figcaption>Image created by Author</figcaption></figure><ol><li>Insert a <strong>Multi-Row Card </strong>and</li><li>Add the following in the <strong>Fields </strong>area:</li></ol><ul><li><strong>Company</strong></li><li><strong>Revenue in USD (MAX)</strong></li></ul><p>3. Apply the following filter:<br> <strong>Rank → Basic Filtering → Select 1</strong></p><h4>Step 2: Format the Card</h4><p>➡️ <strong>Visual Settings</strong></p><ol><li>Turn the following off:</li></ol><ul><li>Category labels</li></ul><p>2. Set the color:</p><ul><li>Cards → Title → Color → #000000</li></ul><p>3. Cards → Accent Bar → Set Color: #8F1238 and Width: 7</p><p>➡️ <strong>General Settings</strong></p><p>Disable the following:</p><ul><li>Title</li><li>Background (Effects)</li></ul><p>Do the same for the other 4 ranks and you are done. Your radial chart is ready!</p><h3>Final Thoughts</h3><p>This radial chart is a great example of how simple data can be transformed into a more engaging and visually appealing story.</p><p>Power BI is not limited to traditional charts. With a bit of creativity, you can design visuals that stand out and make your dashboards more impactful.</p><p>The key takeaway is this: <strong>it’s not only about the data — it’s about how you present it.</strong></p><p>If you found this helpful, feel free to follow for more Power BI and data visualization tutorials.</p><p>For more related articles read these:</p><ul><li><a href="https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-1-55ed6c7e51f4">50 Data Visualization Techniques Every Data Analyst Needs (With Real Use Cases) — Part 1</a></li><li><a href="https://medium.com/microsoft-power-bi/how-to-create-a-star-schema-in-power-bi-step-by-step-guide-7d6a895aa964">How to Create a Star Schema in Power BI (Step-by-Step Guide)</a></li><li><a href="https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-2-84e7b44be0b0">50 Data Visualization Techniques Every Data Analyst Needs (With Real Use Cases) — Part 2</a></li></ul><p><strong>💡 </strong><a href="https://powerbi-masterclass.short.gy/pbi-gpt?utm_source=medium&amp;utm_campaign=pbi-gpt-medium-post-end"><strong>MUST TRY — Power BI GPT — Personal Power BI Learning Coach</strong></a><strong>💡</strong></p><blockquote>Don’t forget to subscribe to</blockquote><blockquote>👉 <a href="https://powerbi-masterclass.short.gy/publication?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=publication-medium-post-end">Power BI Publication</a></blockquote><blockquote>👉 <a href="https://powerbi-masterclass.short.gy/newsletter-medium?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=newsletter-medium-medium-post-end">Power BI Newsletter</a></blockquote><blockquote>and join our Power BI community:</blockquote><p><a href="https://linktr.ee/powerbi.masterclass?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=linktree-medium-post-end&amp;source=post_page-----63d59d50a1eb---------------------------------------">Microsoft Power BI Masterclass | Linktree</a></p><p><strong>Power BI Masterclass Article Classification</strong></p><p><strong>Level:</strong> Beginner</p><p><strong>Category:</strong> Data Visulization</p><p><strong>Tags:</strong> Tutorial, Data Visualization</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=390fe8da920b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/microsoft-power-bi/how-to-create-a-radial-chart-in-power-bi-step-by-step-guide-390fe8da920b">How to Create a Radial Chart in Power BI (Step-by-Step Guide)</a> was originally published in <a href="https://medium.com/microsoft-power-bi">Power BI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[50 Data Visualization Techniques Every Data Analyst Needs (With Real Use Cases) — Part 2]]></title>
            <link>https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-2-84e7b44be0b0?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/84e7b44be0b0</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[power-bi-guide]]></category>
            <category><![CDATA[power-bi]]></category>
            <category><![CDATA[data-visualization]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Thu, 29 Jan 2026 14:51:56 GMT</pubDate>
            <atom:updated>2026-01-30T20:59:16.293Z</atom:updated>
            <content:encoded><![CDATA[<h3>50 Data Visualization Techniques Every Data Analyst Needs (With Real Use Cases) — Part 2</h3><p>In <a href="https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-1-55ed6c7e51f4"><strong>Part 1</strong></a>, we explored visualizations for comparison, correlation, trends, composition, and density.</p><p>In <strong>Part 2</strong>, we’ll continue the list with charts that help you understand <strong>distribution, performance, proportions, behavior, structure, and relationships</strong> along with practical, real-world use cases for each.</p><p>Let’s dive in.</p><p><strong>🎁 </strong><a href="https://powerbi-masterclass.short.gy/learning-database?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=database-medium-post-start"><strong>Get friend links for all of our 1000&gt; Power BI learning articles here</strong></a><strong> 🎁</strong></p><h3>🖈Distribution Visualizations</h3><p>When you want to understand how values are distributed detect spread, or identify outliers then distribution visualizations are the right choice.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*w1sJxI0MsnuuW4lAXFZTWg.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>28. Histogram: </strong>It can help you visualize data shape, center and spread for continuous numerical data by grouping values into bins. <strong>Example: Distribution of order values.</strong></p><p><strong>29. Box Plot:</strong> It is an easy way to display median, quartiles, and outliers. <strong>Example: Delivery time variability by courier.</strong></p><p><strong>30. Violin Plot: </strong>It combines box plot with kernel desnity plot (KDE) to show both summary statistics and distribution shape. <strong>Example: Comparing test score distributions across classes.</strong></p><p><strong>31. Strip Plot: </strong>It is a good way to highlight distribution, clusters, and outliers. Keep in mind that itworks very well for medium and small sized datasets. <strong>Example: Individual transaction amounts.</strong></p><p><strong>32. ECDF Plot (Empirical Cumulative Distribution Function): </strong>It shows every single data point, providing a more detailed view of the data’s cumulative distribution without smoothing. <strong>Example: Percentage of customers spending below a threshold.</strong></p><h3>🖈Gauge / Indicator Visualizations</h3><p>You can use these visualizations when you want to communicate performance against a target or goal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-1_08WCHiubPoa7RUjbQBA.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>33. Gauge Chart: </strong>You can clearly display progress toward a target with this chart. <strong>Example: Monthly revenue progress toward target.</strong></p><p><strong>34. KPI Card: </strong>It is a good way to highlight a single key metric. <strong>Example: Total sales this month.</strong></p><p><strong>35. Scorecard: </strong>If you want to track performance, progress towards a goal all at once. Then scorecard is the right tool. It is actually a card with multiple KPIs in one view. <strong>Example: Sales, profit margin, and churn rate overview.</strong></p><p><strong>36. Progress Bar: </strong>Also, a nice way to track performance. Progress bar is just that, a simple goal tracking. <strong>Example: Project completion percentage.</strong></p><p><strong>37. Traffic Light Indicator: </strong>It uses color-coded status (green/yellow/red) to quickly convey complex data on a status-based performance. <strong>Example: SLA status (On track / At risk / Delayed).</strong></p><h3>🖈Proportional Visualizations</h3><p>Proportional visualizations are often confused with Composition visualizations, which I covered in Part 1. However, there’s an important distinction between them.</p><p>Proportional charts emphasize <strong>size or magnitude</strong>, while Composition charts focus on <strong>how parts contribute to a whole (share)</strong>.</p><p>A simple way to think about it is:</p><ul><li>Composition → One pizza sliced into pieces</li><li>Propotional → Differrent pizzas of different size.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aEkF1GmoMkL0DewZy-eg_w.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>38. Bubble Chart: </strong>Like scatter plot this chart is used to check for relationships between numeric variables. But instead of two variables it uses three. The extra variable (3rd one) is measured with the size of the bubble. <strong>Example: Country population vs Gross Domestic Product (GDP).</strong></p><p><strong>39. Proportional Symbol Map: </strong>You can use it when you want geographic magnitude comparison. <strong>Example: Store revenue by city.</strong></p><p><strong>40. Marimekko Chart: </strong>When you have categorical variables and want to analyse relationships between categories and subcategories simultaneously, this is the right chart for you. <strong>Example: Market share by segment and region.</strong></p><h3>🖈Behavioral Visualizations</h3><p>These charts help you to analyze user or entity behavior, such as retention, engagement, and drop-off.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1I1bMI8EnWTwZE1_T4ZM2g.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>41. Funnel Chart: </strong>It is a great way to visualize conversion stages and highlights drop-offs. <strong>Example: Website conversion funnel.</strong></p><p><strong>42. Cohort Chart: </strong>When you want to show the behavior over time by group. This chart will help you compare how different those groups behave. <strong>Example: Customer retention by signup month.</strong></p><p><strong>43. Retention Curve: </strong>It is a line graph that will help you display user engagement over time. <strong>Example: App usage decay over time.</strong></p><p><strong>44. Heatmap: </strong>It helps you detect activity intensity or frequency patterns. <strong>Example: Login activity by day and hour.</strong></p><p><strong>45. Event Timeline: </strong>Use it when you want to display a sequence of actions, event timeline or milestones. <strong>Example: Login activity by day and hour.</strong></p><h3>🖈Structural Visualizations</h3><p>With structural visualizations you can show hierarchy or system structure. It is a great way to show complex relationships, hierarchies, or spatial arrangements within structured data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*P3oo7S0h_chOWtJiCLPLrA.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>46. Tree Diagram: </strong>It is used to display parent-child relationships. <strong>Example: Product category hierarchy.</strong></p><p><strong>47. Org Chart: </strong>It shows organizational structure. You can display reporting relationships, roles, and hierarchy within a company. <strong>Example: Company organizational structure.</strong></p><p><strong>48. Dendrogram: </strong>It is useful in clustering to show how items group together based on similarity. <strong>Example: Customer clustering results.</strong></p><h3>🖈Relationship Visualizations</h3><p>Use these when you want to show connections, associations, dependencies, or correlations between entities.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AssMmdoRPoafCMsHKq3A6A.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>49. Network Graph: </strong>With network graphs you can show nodes and relationships as links. <strong>Example: Social media interactions between users.</strong></p><p><strong>50. Chord Diagram: </strong>A great way to present flow and interconnections between categories. <strong>Example: Trade flows between countries.</strong></p><h3>Final Thoughts</h3><p>Choosing the right chart isn’t about using what looks fancy, it’s about <strong>matching the visual to the question you’re trying to answer</strong>.</p><p>With these 50 visualization techniques across both parts, you now have a practical toolkit for almost any analytical scenario.</p><p>If there is any chart you would like me to cover in more depth with step-by-step examples (Power BI, Python, or Tableau), leave a comment, I’d love to create a detailed guide for it.</p><p>If this guide helped you think differently about chart selection, consider clapping 👏 and follow for more related content.</p><p><strong>💡 </strong><a href="https://powerbi-masterclass.short.gy/pbi-gpt?utm_source=medium&amp;utm_campaign=pbi-gpt-medium-post-end"><strong>MUST TRY — Power BI GPT — Personal Power BI Learning Coach</strong></a><strong>💡</strong></p><blockquote>Don’t forget to subscribe to</blockquote><blockquote>👉 <a href="https://powerbi-masterclass.short.gy/publication?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=publication-medium-post-end">Power BI Publication</a></blockquote><blockquote>👉 <a href="https://powerbi-masterclass.short.gy/newsletter-medium?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=newsletter-medium-medium-post-end">Power BI Newsletter</a></blockquote><blockquote>and join our Power BI community:</blockquote><p><a href="https://linktr.ee/powerbi.masterclass?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=linktree-medium-post-end&amp;source=post_page-----cae7b7ef9d6c---------------------------------------">Microsoft Power BI Masterclass | Linktree</a></p><p><strong>Power BI Masterclass Article Classification</strong></p><p><strong>Level:</strong> Beginner</p><p><strong>Category:</strong> Data Visualization</p><p><strong>Tags:</strong> Tutorial, Data Visualization</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=84e7b44be0b0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-2-84e7b44be0b0">50 Data Visualization Techniques Every Data Analyst Needs (With Real Use Cases) — Part 2</a> was originally published in <a href="https://medium.com/microsoft-power-bi">Power BI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[50 Data Visualization Techniques Every Data Analyst Needs (With Real Use Cases) — Part 1]]></title>
            <link>https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-1-55ed6c7e51f4?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/55ed6c7e51f4</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[power-bi]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[power-bi-guide]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Wed, 21 Jan 2026 15:59:39 GMT</pubDate>
            <atom:updated>2026-06-18T05:25:38.923Z</atom:updated>
            <content:encoded><![CDATA[<h3>50 Data Visualization Techniques Every Data Analyst Needs (With Real Use Cases) — Part 1</h3><p>When I first started building dashboards, one of my biggest challenges was to choose the right visual to represent my data.</p><p><strong>Selecting the right visualization is just as important as the data itself. </strong>A great chart, as you might already know, it’s what makes insights obvious. A poor one just hides the story.</p><p>That’s why I decided to write this guide with the <strong>50 most commonly used data visualization techniques</strong>, organized by <strong>purpose</strong>, not tools. The goal, here, is to help analysts choose visuals based on <em>what they want to communicate</em>, not just what a tool offers.</p><p>So, whether you’re building dashboards in Power BI, Tableau, or Python, I am sure that this list will help you consistently pick the <em>right visual for the right question</em>.</p><p><strong>🎁 </strong><a href="https://powerbi-masterclass.short.gy/learning-database?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=database-medium-post-start"><strong>Get friend links for all of our 1500&gt; Power BI learning articles here</strong></a><strong> 🎁</strong></p><h3>🖈Comparison Visualizations</h3><p>These visualizations will help you to show the differences or similarities between categories. So, use them when you want to compare values across categories.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*by08xvzZmpxXyqBtXwqd0w.jpeg" /><figcaption>Image created by Author</figcaption></figure><ol><li><strong>Bar Chart: </strong>It compares values across categories. It’s the most common and the most effective one. <strong>Example: Show a comparison of total sales by product category.</strong></li></ol><p><strong>2. Column Chart: </strong>It is the same as a bar chart but vertical. You can use it to compare items and it is also great for time-based comparisons. <strong>Example: It can show a monthly revenue comparison across regions or different products.</strong></p><p><strong>3. Grouped Bar Chart: </strong>If you want to compare multiple metrics across categories, this is the right chart for you. <strong>Example: You can compare sales with profit for each store or sales between female and male employees.</strong></p><p><strong>4. Lollipop Chart: </strong>It is an alternative to a bar chart. If you have many categories that have similar-sized values, this chart would be cleaner and less cluttered than a bar chart. <strong>Example: You can set side by side employee performance scores.</strong></p><p><strong>5. Dot Plot: </strong>It is a good way to highlight small differences. Use it if you want to compare single quantitative variables. <strong>Example: Compare average delivery times by courier.</strong></p><p><strong>6. Bullet Chart: </strong>It is the chart that can easily help you compare actual vs target performance. <strong>Example: You can display actual sales vs target sales per salesperson.</strong></p><p><strong>7. Slope Chart:</strong> It is a minimalistic chart that compares changes between two specific points in time. <strong>Example: Customer satisfaction before and after a product update.</strong></p><h3>🖈Correlational Visualizations</h3><p>When you want to understand relationships between variables, this type of visualisations are the most suitable for this job.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5zpmBF-VbF-9pOSkTwTsBw.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>8. Scatter Plot: </strong>It’s an effective way to detect correlations, trends and patterns between two numerical variables. <strong>Example: Advertising spend vs revenue.</strong></p><p><strong>9. Bubble Chart: </strong>It does the same thing as scatter plot but you can use a third variable encoded as size. <strong>Example: Sales vs profit, bubble size = market size.</strong></p><p><strong>10. Correlation Matrix: </strong>It is a table that can visualize correlation across multiple variables. <strong>Example: It can show you the correlation between customer age, income and spend all at one table.</strong></p><p><strong>11. Pair Plot: </strong>It is a matrix of graphs that enables the visualization of the relationship between each pair of variables. <strong>Example: You can explore relationships between multiple marketing metrics.</strong></p><p><strong>12. Hexbin Plot: </strong>It can visualize density-based correlation for large datasets. So, if you want to visualise the relationship between two variables but the dataset is too large, hexbin plot is the the chart for this job. <strong>Example: Website session duration vs page view.</strong></p><h3>🖈Trend Visualizations</h3><p>Trend visualizations are suitable if you want to show how values change over time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JfEN_-qTKnjniYjVbOUttg.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>13. Line Chart: </strong>You can use it when you want to visualize continuous time trends. <strong>Example: Daily active users or monthly revenue over a given period of time.</strong></p><p><strong>14. Area Chart: </strong>It is a nice way to emphasize volume and magnitude of change over time. <strong>Example: Total revenue growth or company’s website traffic change over time.</strong></p><p><strong>15. Stacked Area Chart: </strong>It allows you to display contributions from different category values along with their totals. <strong>Example: Revenue contribution by product line over time.</strong></p><p><strong>16. Step Line Chart: </strong>It is similar to a line chart but you can use it to show the changes that occur at irregular intervals. SO, it is suitable to highlight discrete changes. <strong>Example: Price changes applied at fixed intervals.</strong></p><p><strong>17. Sparkline: </strong>They are tiny trend chart that can be used in a table. If you want something extremely small to fit in a text or a cell, this is the chart for you. <strong>Example: Sales trend next to each product in a table.</strong></p><p><strong>18. Rolling Average Line: </strong>It can smooths noisy time-series data. It works by calculating the average of a specific number of data points and plotting this value as a continuous line on a chart. <strong>Example: 7-day moving average of website traffic.</strong></p><h3>🖈Composition Visualizations</h3><p>When you want to show how individual categories contribute to a total these type of charts work like a charm.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iGg6Zd5SYmcnU9Ng47QWYg.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>19. Pie Chart: </strong>It can show in a simple way a part-to-whole comparison (use with 2–4 categories). <strong>Example: Market share by brand.</strong></p><p><strong>20. Donut Chart: </strong>Cleaner alternative to pie charts. <strong>Example: Expense breakdown for a department.</strong></p><p><strong>21. Stacked Bar Chart: </strong>You can use it to compare composition across categories. <strong>Example: Sales split by channel for each region.</strong></p><p><strong>22. 100% Stacked Bar: </strong>Compare proportions, not values. It visually shows the percentage of the whole that each subcategory constitutes, with the whole always being 100%, for different series. <strong>Example: Customer segment proportions by country.</strong></p><p><strong>23. Tree Map: </strong>Use them when you want to display hierarchical part-to-whole relationships. <strong>Example: Revenue contribution by product hierarchy.</strong></p><p><strong>24. Sunburst Chart: </strong>It is ideal for displaying part-to-whole relationships across multiple categories like organizational structures or product breakdowns. So, if you want multi-level composition this is the chart for you. <strong>Example: Website navigation paths by page level.</strong></p><h3>🖈Density Visualizations</h3><p>They are perfect if you want to visualize the distribution of numerical, continuous data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bcLyWjlgJJ8W1dFLx4mejw.jpeg" /><figcaption>Image created by Author</figcaption></figure><p><strong>25. Kernel Density Plot (KDE):</strong> Smoothed distribution curve. <strong>Example: Distribution of customer purchase amounts.</strong></p><p><strong>26. Density Heatmap: </strong>It shows concentration of data points. <strong>Example: User clicks across a webpage.</strong></p><p><strong>27. Ridgeline Plot: </strong>It helps you to compare distributions across categories. <strong>Example: Salary distribution by department.</strong></p><h3>Final Thoughts</h3><p>Every chart in this guide serves a specific purpose: comparing values, showing trends, revealing relationships, or explaining composition. When you start with <em>what you want your audience to understand</em>, the right visual often becomes obvious.</p><p>This is <strong>Part 1</strong> of the series. In the next part, we’ll explore <strong>distribution, behavioral, relational, structural, and indicator visualizations</strong>, along with real-world use cases.</p><p>If this guide helped you think differently about chart selection, consider clapping 👏 and following for the next part of the series.</p><p>Part 2:</p><p><a href="https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-2-84e7b44be0b0">https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-2-84e7b44be0b0</a></p><p><strong>💡 </strong><a href="https://powerbi-masterclass.short.gy/pbi-gpt?utm_source=medium&amp;utm_campaign=pbi-gpt-medium-post-end"><strong>MUST TRY — Power BI GPT — Personal Power BI Learning Coach</strong></a><strong>💡</strong></p><blockquote>Don’t forget to subscribe to</blockquote><blockquote>👉 <a href="https://powerbi-masterclass.short.gy/publication?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=publication-medium-post-end">Power BI Publication</a></blockquote><blockquote>👉 <a href="https://powerbi-masterclass.short.gy/newsletter-medium?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=newsletter-medium-medium-post-end">Power BI Newsletter</a></blockquote><blockquote>and join our Power BI community:</blockquote><p><a href="https://linktr.ee/powerbi.masterclass?utm_source=medium&amp;utm_medium=medium-post&amp;utm_campaign=linktree-medium-post-end&amp;source=post_page-----d5abefd2dd10---------------------------------------">Microsoft Power BI Masterclass | Linktree</a></p><p><strong>Power BI Masterclass Article Classification</strong></p><p><strong>Level:</strong> Beginner</p><p><strong>Category:</strong> Data Visualization</p><p><strong>Tags:</strong> Guide, Data Visualization</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=55ed6c7e51f4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/microsoft-power-bi/50-data-visualization-techniques-every-data-analyst-needs-with-real-use-cases-part-1-55ed6c7e51f4">50 Data Visualization Techniques Every Data Analyst Needs (With Real Use Cases) — Part 1</a> was originally published in <a href="https://medium.com/microsoft-power-bi">Power BI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Detect Hidden Data Issues in Power BI with a Python-Powered Profiling Dashboard (Part 3)]]></title>
            <link>https://medium.com/codex/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-3-e8aa466f6640?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/e8aa466f6640</guid>
            <category><![CDATA[python]]></category>
            <category><![CDATA[power-bi-tutorials]]></category>
            <category><![CDATA[dashboard-design]]></category>
            <category><![CDATA[data-cleaning]]></category>
            <category><![CDATA[data-profiling]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Thu, 08 Jan 2026 07:41:05 GMT</pubDate>
            <atom:updated>2026-01-08T11:05:19.491Z</atom:updated>
            <content:encoded><![CDATA[<p>In <a href="https://medium.com/codex/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-1-6508a43df551">Part 1</a>, we explored why data profiling matters and how to build a Python script that will help us to detect hidden data quality issues inside Power BI.</p><p>In <a href="https://medium.com/codex/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-2-8b40280e9f15">Part 2</a>, we focused on how to build an overview of data quality using KPIs like completeness, uniqueness, and accuracy.</p><p>Now, in Part 3, we move from the big picture to the details. We are going to:</p><ul><li>Build<strong> dynamic column-level visuals.</strong></li><li>Use <strong>bookmarks </strong>to switch between columns.</li><li>Analyze <strong>distributions, statistics, and data issues per column.</strong></li></ul><p>By the end we will have a fully interactive data profiling dashboard that will help us to explore data quality column by column.</p><p>The final dashboard would look like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vK6yACRWphSFKMqsZcItlg.gif" /></figure><h3>Dashboard Structure</h3><p>In this part of our dashboard we wanted to focus on the quality of each column seperately.</p><p>To achieve this, we structured this area into three different logical areas.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bpTmyYulNrRIB8Lw9lcwEA.jpeg" /><figcaption>Image created by Author</figcaption></figure><h4>1. Column KPIs</h4><p>At the left top corner of this area we are going to include 3 KPIs that update dynamically based on the selected column.</p><ul><li><strong>Column Type</strong> (numeric or object)</li><li><strong>% of Missing Values</strong></li><li><strong>Number of Duplicate Rows</strong></li></ul><h4>2. Distribution Chart</h4><p>At the bottom left corner, we will add a distribution chart that is going to change behavior depending on the column type:</p><ul><li><strong>Object columns</strong> → frequency per category</li><li><strong>Numeric columns</strong> → value distribution using custom bins</li></ul><h4>3. Detail Tables</h4><p>At the right corner we will show two tables that change dynamically:</p><p><strong>For object columns:</strong></p><ul><li>Summary statistics</li><li>Text consistency profile</li></ul><p><strong>For numeric columns:</strong></p><ul><li>Summary statistics</li><li>Outlier analysis</li></ul><h3>Create the Bookmarks</h3><p>Power BI does not allow us to use a single visual to dynamically switch between completely different tables, so <strong>bookmarks are the best solution</strong> here.</p><p>We are going to make 8 bookmarks (one for each column):</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NxDPqY5WMswmZNByf58cKw.jpeg" /></figure><p><strong>Steps:</strong></p><ol><li>Go to the View → Choose Bookmarks</li><li>Add one new bookmark per column</li><li>Rename them clearly: <strong>Item, Location, Payment Method, Price Per Unit, Quantity, Total Spent, Transaction Date, Transaction ID</strong></li></ol><p>Each bookmark will show the KPIs, distribution and tables related to each column.</p><h4>Create Column Selection Buttons</h4><p>As you can see in the image above we have created buttons for each column.</p><p><strong>Steps:</strong></p><ol><li>Insert a <strong>button</strong></li><li>Assign it to the corresponding bookmark</li><li>Label it clearly (e.g., <em>Quantity</em>, <em>Item</em>, <em>Total Spent</em>)</li></ol><p>This will create a smooth, app-like experience.</p><h3>Build KPIs</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tQYYNfaNOQgyRRWHHQyVpQ.jpeg" /></figure><h4>1. Column Type</h4><ul><li>Visualizations → Choose Card → Fields: Data Type (From Profile Table)</li></ul><p>This will tell us instantly whether we analyze a numeric or a categorical column.</p><h4>2. %_Missing_Values</h4><ul><li>Visualizations → Choose Card → Fields: Missing_% (From Profile Table)</li></ul><p>This will help us check if the column needs cleaning or imputation.</p><h4>3. Duplicates</h4><ul><li>Visualizations → Choose Card → Fields: Duplicate_Rows (From Duplicates_Summary Table)</li></ul><p>It gives us the number of duplicates in our dataset. Even a small number of duplicates can distort aggregations and trends, so keep that in mind.</p><h3>Distribution Charts (Numeric vs Object Columns)</h3><h4>Why not use Power BI auto-bins?</h4><p>Power BI auto-bins are great for exploration. But what we want here is to use <strong>rule-based bins</strong> to highlight negative values, surface extreme ranges and make data issues obvious.</p><p>Distribution chart:</p><ul><li>For Object Columns: We will show the quantity of each category.</li><li>For Numerical Columns: We will create bins with DAX measures.</li></ul><p>Note: I am leaving Transaction Date and Transaction ID columns out of this area as I don’t want to analyse them further.</p><h4>Numeric Column Bins (DAX)</h4><p><strong>Price Per Unit Bin:</strong></p><pre>Price_Per_Unit_Bin = <br>SWITCH(<br>    TRUE(),<br>    &#39;Dataset&#39;[Price Per Unit] &lt; 0, &quot;negative value&quot;,    <br>    &#39;Dataset&#39;[Price Per Unit] &lt; 10, &quot;0 - 10&quot;,<br>    &#39;Dataset&#39;[Price Per Unit] &lt; 20, &quot;10 - 20&quot;,<br>    &#39;Dataset&#39;[Price Per Unit] &lt; 30, &quot;20 - 30&quot;,<br>    &#39;Dataset&#39;[Price Per Unit] &gt; 30, &quot;+30&quot;, <br>    &quot; &quot;<br>)</pre><p><strong>Quantity Bin:</strong></p><pre>Quantity_Bin = <br>SWITCH(<br>    TRUE(),<br>    &#39;Dataset&#39;[Quantity] &lt; 0, &quot;negative value&quot;,    <br>    &#39;Dataset&#39;[Quantity] &lt; 2, &quot;1 - 2&quot;,<br>    &#39;Dataset&#39;[Quantity] &lt; 4, &quot;3 - 4&quot;,<br>    &#39;Dataset&#39;[Quantity] &gt; 4, &quot;5+&quot;, <br>    &quot; &quot;<br>)</pre><p><strong>Total Spent Bin:</strong></p><pre>Total_Spent_Bin = <br>SWITCH(<br>    TRUE(),<br>    &#39;Dataset&#39;[Total Spent] &lt; 0, &quot;negative value&quot;,    <br>    &#39;Dataset&#39;[Total Spent] &lt; 100, &quot;0 - 100&quot;,<br>    &#39;Dataset&#39;[Total Spent] &lt; 200, &quot;100 - 200&quot;,<br>    &#39;Dataset&#39;[Total Spent] &lt; 300, &quot;200-300&quot;,<br>    &#39;Dataset&#39;[Total Spent] &gt; 300, &quot;+300&quot;,<br>    &quot; &quot;<br>)</pre><h4>Create the Distribution Visual</h4><p><strong>Steps:</strong></p><ol><li>Go to visualization pane</li><li>Select Stacked column chart</li><li>Use:</li></ol><ul><li>Axis → Bin column (or category column)</li><li>Values → Row count or relevant measure</li></ul><p>Each bookmark that we created above will control which bin or category is active.</p><h3>Table Visuals</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3cD8hG-Il1iqtCve9HCI6Q.jpeg" /></figure><p>This area will visualize the tables that we have created with Python script:</p><ul><li>For Object Columns: <strong>Summary_Obj</strong> and <strong>Text_Profile </strong>tables</li><li>For Numerical Columns: <strong>Summary_Num </strong>and <strong>Outlier_Profile </strong>tables</li></ul><h4>Create the tables</h4><p><strong>Steps:</strong></p><ol><li>Go to visualization pane</li><li>Select matrix</li></ol><p>Make sure that each bookmark toggles the appropriate tables into view.</p><h3>Final Thoughts</h3><p>With Part 3, we have completed our journey to create a Data Profiling Dashboard from raw data.</p><p>You now have:</p><ul><li>Automated Python-based profiling</li><li>High-level data quality KPIs</li><li>Column-level diagnostics</li><li>Bookmark-driven interactivity</li></ul><p>If you enjoyed this series, feel free to 👏 clap and follow for more advanced Power BI, Python, and data quality content.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e8aa466f6640" width="1" height="1" alt=""><hr><p><a href="https://medium.com/codex/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-3-e8aa466f6640">How to Detect Hidden Data Issues in Power BI with a Python-Powered Profiling Dashboard (Part 3)</a> was originally published in <a href="https://medium.com/codex">CodeX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Detect Hidden Data Issues in Power BI with a Python-Powered Profiling Dashboard (Part 2)]]></title>
            <link>https://medium.com/codex/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-2-8b40280e9f15?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/8b40280e9f15</guid>
            <category><![CDATA[power-bi]]></category>
            <category><![CDATA[data-preprocessing]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[power-bi-tutorials]]></category>
            <category><![CDATA[data-cleaning]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Mon, 08 Dec 2025 06:58:05 GMT</pubDate>
            <atom:updated>2025-12-10T06:45:47.270Z</atom:updated>
            <content:encoded><![CDATA[<p>In Part 1, we talked about the importance of Data Profiling and how to create a Python script to automatically detect missing values, outliers, duplicates and more.</p><p>Now in Part 2 we will use that Python script to build an interactive Data Quality Dashboard in Power BI</p><p>This article includes:</p><ul><li>Preparation of our dataset</li><li>Running the Python script correctly</li><li>Generating tables</li><li>Designing the overview section (Building KPIs and gauges to check the overall health of our dataset)</li></ul><p>The Final Dashboard:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vK6yACRWphSFKMqsZcItlg.gif" /></figure><h3>Prepare Dataset</h3><p>(Dataset used for this example: <a href="https://www.kaggle.com/datasets/ahmedmohamed2003/cafe-sales-dirty-data-for-cleaning-training/data">Cafe Sales</a> from Kaggle)</p><p>Open our dataset in <strong>Transform Data</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xmBc7PoTcJnL_3j2Jf3GsQ.jpeg" /><figcaption>Image created by Author</figcaption></figure><p>As we can check by clicking in the arrow of every column the dataset has blank spaces and “ERROR” or “UNKNOWN” strings. These inconsistencies will break your profiling logic unless you normalize them. So the first thing that we have to do is to replace blank, ERROR and UNKNOWN values with NaN.</p><p>Select<strong> all</strong> the columns → Click on <strong>Replace Values </strong>→</p><p>Replace these values one by one:</p><ul><li>Value to Find : <strong>blank</strong> Replace with : <strong>NaN</strong></li><li>Value to Find : <strong>ERROR </strong>Replace with : <strong>NaN</strong></li><li>Value to Find : <strong>UNKNOWN </strong>Replace with : <strong>NaN</strong></li></ul><p>This will make our missing data consistent, and ensure Python interprets them properly.</p><h3>Run the Python script</h3><p>Now that our dataset is ready to process, let’s run our Python script.</p><p><strong>Duplicate </strong>your Dataset → Go to <strong>Transform </strong>→ Click on <strong>Run Python Script</strong> → <strong>Paste </strong>the <strong>Python script </strong>from Part 1 (included below)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SGgCAorou5ZeYGTWqAdnuA.jpeg" /><figcaption>Image created by Author</figcaption></figure><pre># --- Data Profiling Script for Power BI ---<br><br># --- Import Required Libraries ---<br>import pandas as pd<br>import numpy as np<br><br># --- 1. BASIC OVERVIEW ---<br><br>profile = pd.DataFrame({<br>    &quot;Column&quot;: dataset.columns,<br>    &quot;Data_Type&quot;: dataset.dtypes.astype(str),<br>    &quot;Missing_Values&quot;: dataset.isnull().sum(),<br>    &quot;Missing_%&quot;: (dataset.isnull().mean() * 100).round(2),<br>    &quot;Unique_Values&quot;: dataset.nunique(),<br>    &quot;Total_Values&quot;: len(dataset),<br>    &quot;Total_Rows&quot;: len(dataset)<br>})<br><br># --- 2. TEXT CONSISTENCY CHECK (for object columns) ---<br><br># Select object-type columns with at least 1 non-null value<br>text_cols = [<br>    col for col in dataset.select_dtypes(include=[&#39;object&#39;]).columns<br>    if dataset[col].notna().sum() &gt; 0<br>]<br><br>text_report = []<br><br>for col in text_cols:<br>    sample = dataset[col].dropna().astype(str)<br>    total_rows = len(sample)<br>    if total_rows == 0:<br>        continue<br><br>    # Count casing patterns<br>    all_upper = (sample == sample.str.upper()).sum()<br>    all_lower = (sample == sample.str.lower()).sum()<br>    title_case = (sample == sample.str.title()).sum()  <br><br>    # Detect leading/trailing spaces<br>    spaces = sample[sample.str.strip() != sample].count()<br><br>    # Calculate uniformity (percentage of dominant style)<br>    dominant_case = max(all_upper, all_lower, title_case)<br>    case_uniformity = round((dominant_case / total_rows) * 100, 2)<br><br>    # Additional consistency checks<br>    numeric_like = sample.str.isnumeric().sum()<br>    contains_digits = sample.str.contains(r&quot;\d&quot;).sum()<br><br>    text_report.append({<br>        &quot;Column&quot;: col,<br>        &quot;Total_Values&quot;: total_rows,<br>        &quot;Unique_Values&quot;: sample.nunique(),<br>        &quot;All_Uppercase_Count&quot;: int(all_upper),<br>        &quot;All_Lowercase_Count&quot;: int(all_lower),<br>        &quot;Title_Case_Count&quot;: int(title_case),<br>        &quot;Leading_Trailing_Spaces&quot;: int(spaces),<br>        &quot;Numeric_Like_Count&quot;: int(numeric_like),<br>        &quot;Contains_Digits_Count&quot;: int(contains_digits),<br>        &quot;Case_Uniformity_%&quot;: case_uniformity<br>    })<br><br>text_profile = pd.DataFrame(text_report)<br><br># --- 3. OUTLIER DETECTION (IQR method) ---<br><br>numeric_cols = dataset.select_dtypes(include=np.number).columns<br><br>outlier_report = []<br><br>for col in numeric_cols:<br>    q1 = dataset[col].quantile(0.25)<br>    q3 = dataset[col].quantile(0.75)<br>    iqr = q3 - q1<br>    lower_bound = q1 - 1.5 * iqr<br>    upper_bound = q3 + 1.5 * iqr<br><br>    outliers = dataset[(dataset[col] &lt; lower_bound) | (dataset[col] &gt; upper_bound)][col]<br><br>    outlier_report.append({<br>        &quot;Column&quot;: col,<br>        &quot;Min&quot;: dataset[col].min(),<br>        &quot;Max&quot;: dataset[col].max(),<br>        &quot;Lower_Bound&quot;: round(lower_bound, 2),<br>        &quot;Upper_Bound&quot;: round(upper_bound, 2),<br>        &quot;Outlier_Count&quot;: len(outliers),<br>        &quot;Outlier_%&quot;: round(len(outliers) / len(dataset) * 100, 2)<br>    })<br><br>outlier_profile = pd.DataFrame(outlier_report)<br><br># --- 4. DUPLICATE CHECK ---<br><br># Check for full row duplicates<br>total_duplicates = dataset.duplicated().sum()<br><br>duplicate_rows = dataset[dataset.duplicated(keep=False)] <br><br>duplicates_summary = pd.DataFrame({<br>    &quot;Total_Rows&quot;: [len(dataset)],<br>    &quot;Duplicate_Rows_Count&quot;: [len(duplicate_rows)],<br>    &quot;Unique_Rows_Count&quot;: [len(dataset) - total_duplicates],<br>    &quot;Duplicates_%&quot;: [round((total_duplicates / len(dataset)) * 100, 2)]<br>})<br><br># --- 5. ACCURACY RULES ---<br><br>accuracy_rules = []<br><br># Rule 1: Price Per Unit &gt; 0<br>price_valid = (dataset[&#39;Price Per Unit&#39;] &gt; 0).sum()<br>price_total = dataset[&#39;Price Per Unit&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;:&#39;Price Per Unit&#39;,<br>    &#39;Valid_Count&#39;: price_valid,<br>    &#39;Total_Count&#39;: price_total,<br>    &#39;Accuracy_%&#39;: round(price_valid / price_total * 100, 2)<br>})<br><br># Rule 2: Total Spent &gt; 0<br>Total_Spent_valid = (dataset[&#39;Total Spent&#39;] &gt; 0).sum()<br>Total_Spent_total = dataset[&#39;Total Spent&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;: &#39;Total Spent&#39;,<br>    &#39;Valid_Count&#39;: Total_Spent_valid,<br>    &#39;Total_Count&#39;: Total_Spent_total,<br>    &#39;Accuracy_%&#39;: round(Total_Spent_valid / Total_Spent_total * 100, 2)<br>})<br><br># Rule 3: Quantity between 1 and 100<br>qty_valid = ((dataset[&#39;Quantity&#39;] &gt;= 1) &amp; (dataset[&#39;Quantity&#39;] &lt;= 100)).sum()<br>qty_total = dataset[&#39;Quantity&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;: &#39;Quantity&#39;,<br>    &#39;Valid_Count&#39;: qty_valid,<br>    &#39;Total_Count&#39;: qty_total,<br>    &#39;Accuracy_%&#39;: round(qty_valid / qty_total * 100, 2)<br>})<br><br>accuracy = pd.DataFrame(accuracy_rules)<br><br># --- 6. SUMMARY STATISTICS (numeric &amp; object)---<br><br>summary_stats_obj = dataset.describe(include=&#39;object&#39;).T.reset_index().rename(columns={&#39;index&#39;: &#39;Column&#39;})<br><br>summary_stats_num = dataset.describe(include=np.number).T.reset_index().rename(columns={&#39;index&#39;: &#39;Column&#39;})<br><br># --- 7. OUTPUT TABLES FOR POWER BI ---<br># You can duplicate and rename in Power BI to show as separate tables<br>profile, text_profile, outlier_profile, duplicate_rows, duplicates_summary, accuracy, summary_stats_obj, summary_stats_num</pre><p>If you want to learn more about the above Python script you can check Part 1:</p><p><a href="https://medium.com/@ritaaggelou/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-1-6508a43df551">How to Detect Hidden Data Issues in Power BI with a Python-Powered Profiling Dashboard (Part 1)</a></p><p>Power BI will execute the Python code and generate <strong>8 tables</strong>:</p><ul><li>Accuracy</li><li>Dataset</li><li>Duplicates Summary</li><li>Duplicate Rows</li><li>Outlier Profile</li><li>Profile</li><li>Summary Stats (Numeric)</li><li>Summary Stats (Object)</li><li>Text Profile</li></ul><p>You can now duplicate each table in Power Query and rename them logically.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*H-phy3UZvSk7nxMK8t5i_w.jpeg" /></figure><p><strong>Home </strong>→ Click on <strong>Close and Apply.</strong></p><h3>Correct Data Types</h3><p>Before we start creating our dashboard make sure that all the columns from your new tables have the correct data type.</p><ul><li>text profiling tables use <strong>text columns</strong></li><li>numeric profiling tables use <strong>whole/decimal numbers</strong></li><li>summary stats have <strong>proper numeric formatting</strong></li></ul><p>If we do not fix the column types we will have broken visuals laters. So, don’t skip this step, it’s an important one.</p><h3>Build the Data Quality Overview Section</h3><p>The overview part of our Data Profile Dashboard has:</p><ul><li>3 KPIs (Total Rows, Unique Rows and Duplicates).</li><li>3 Gauge charts (Completeness, Uniqueness and the Accuracy)</li><li>1 Overall Data Quality Score</li></ul><p>This is the first thing that our viewers will notice and it talks about the overall health of our dataset, so clarity is the key here.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5C6m2gzP8fh8kI8EbjZubA.jpeg" /><figcaption>Image created by Author</figcaption></figure><h4>1. KPI: Total Rows</h4><p>Create a measure with this Dax:</p><pre>Total Rows = COUNTROWS(&#39;Dataset&#39;) </pre><ul><li>Visualizations → Choose Card → Fields: Total Rows</li></ul><h4>2. KPI: Unique Rows</h4><p>Create a measure with this Dax:</p><pre>Unique Rows = DISTINCTCOUNT(&#39;Dataset&#39;[Transaction ID])</pre><ul><li>Visualizations → Choose Card → Fields: Unique Rows</li></ul><h4>3. KPI: Duplicates</h4><p>Create a measure with this Dax:</p><pre>No of Duplicates = COUNTROWS(&#39;Duplicates&#39;)</pre><ul><li>Visualizations → Choose Card → Fields: No of Duplicates</li></ul><h4>4. Completeness Gauge (Missing Data)</h4><p>Completeness measures all the missing values in a dataset.</p><p>To create the completeness measure:</p><ul><li>create the <strong>Average Missing % measure</strong>:</li></ul><pre>Average Missing % = AVERAGEX(<br>    &#39;Profile&#39;, DIVIDE(&#39;Profile&#39;[Missing_Values], &#39;Profile&#39;[Total_Values])<br>)</pre><ul><li>and then create the <strong>Completeness measure</strong>:</li></ul><pre>Completeness = 1 - [Average Missing %]</pre><ul><li>Visualizations → Choose Gauge → Value: Completeness</li></ul><p>This visual will help us to have a quick view about how complete is our dataset.</p><h4>5. Uniqueness Gauge (Duplicate Ratio)</h4><p>Uniqueness is the measure that will show us the duplicates of our dataset.</p><ul><li>create the <strong>Uniqueness measure</strong>:</li></ul><pre>Uniqueness = DIVIDE([Unique Rows], [Total Rows])</pre><ul><li>Visualizations → Choose Gauge → Value: Uniqueness</li></ul><p>This visual will tell us how much of our dataset is distinct.</p><h4>6. Accuracy Gauge (Business-Rule Validation)</h4><p>Accuracy is the measure that will help us check if our data reflect real-world values.</p><ul><li>create the <strong>Accuracy measure</strong>:</li></ul><pre>Accuracy = <br>VAR OverallAccuracy = SUM(&#39;Accuracy_Num&#39;[Accuracy_%])<br>VAR ColumnCount = COUNTROWS(&#39;Accuracy_Num&#39;)*100<br>RETURN<br>DIVIDE(OverallAccuracy, ColumnCount)</pre><ul><li>Visualizations → Choose Gauge → Value: Accuracy</li></ul><h4>7. Overall Data Quality Score</h4><p>Finally, we will combine the above measures (completeness, uniqueness, and accuracy) into one metric.</p><ul><li>create the <strong>Data Quality Score measure</strong>:</li></ul><pre>Data Quality Score = <br>DIVIDE([Uniqueness]+[Accuracy]+[Completeness], 3)</pre><ul><li>Visualizations → Choose Card → Fields: Data Quality Score</li></ul><p><strong>Tip:</strong> The formula treats all three dimensions equally but you can weight them differently depending on your use case.</p><h3>Final Thoughts</h3><p>At this stage, we’ve built the foundational layer of our Data Profiling Dashboard. The viewer now has a clear, high-level understanding of the overall quality of our dataset.</p><p>But Data Profiling doesn’t stop there. We have to examine the quality of each column individually. <strong>So, how can we visualize text issues, numeric outliers, and summary statistics column by column without overwhelming the report?</strong></p><p>That is part of my next article.</p><h3>What’s next? (Part 3)</h3><p>In Part 3, the final part of this series we are going to:</p><ul><li>Build Dynamic column-level visuals</li><li>Create bookmarks to switch between charts per column</li></ul><p>🌟 If you enjoyed this article dont’ forget to 👏 clap and follow for more related content!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8b40280e9f15" width="1" height="1" alt=""><hr><p><a href="https://medium.com/codex/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-2-8b40280e9f15">How to Detect Hidden Data Issues in Power BI with a Python-Powered Profiling Dashboard (Part 2)</a> was originally published in <a href="https://medium.com/codex">CodeX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Detect Hidden Data Issues in Power BI with a Python-Powered Profiling Dashboard (Part 1)]]></title>
            <link>https://medium.com/codex/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-1-6508a43df551?source=rss-79a3418e04a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/6508a43df551</guid>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[data-cleaning]]></category>
            <category><![CDATA[power-bi]]></category>
            <category><![CDATA[data-processing]]></category>
            <dc:creator><![CDATA[Rita Angelou]]></dc:creator>
            <pubDate>Wed, 03 Dec 2025 08:23:29 GMT</pubDate>
            <atom:updated>2025-12-12T06:38:29.225Z</atom:updated>
            <content:encoded><![CDATA[<p>Have you ever built a dashboard where the results just didn’t look right? Maybe a total seemed off, a trend didn’t make sense, or something simply wasn’t adding up.</p><p>Cleaning your data before building a model or dashboard is essential, but there is an important step that comes <strong>before</strong> cleaning:<br> 👉 <strong>Checking the quality of your data.</strong></p><p>If your data is low quality, it doesn’t matter how much you clean. The insights will still be unreliable.</p><p>This is where <strong>data profiling </strong>comes in. It’s the process of examining your raw dataset to uncover hidden problems (like missing values, wrong data types, inconsistent formatting, duplicated rows, and outliers).</p><p>At this point you might think: <strong>“But Power BI already has built-in profiling tools… why bother doing this manually?”</strong></p><p>Well, yes and no. Power BI’s Column Quality, Column Distribution, and Column Profile are useful, but it doesn’t let you run deep text consistency checks, identify case-format issues, analyze outliers with IQR, or generate custom accuracy rules.</p><p><strong>That’s where Python steps in to rescue the process.</strong></p><p>By combining Python with Power BI, you can automate advanced data diagnostics and turn them into visual, interactive, and actionable insights.</p><p>This article (Part 1) covers:</p><ul><li>Why Data Profiling matters</li><li>How I built a full Data Profiling Dashboard inside Power BI (using Cafe Sales dataset from Kaggle)</li><li>How the Python script works line by line</li></ul><p>In Part 2 I’ll show you how to turn those tables into an interactive Power BI dashboard (DAX, bookmarks, tooltips, conditional formatting, and layout tips).</p><p>Final Dashboard look:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vK6yACRWphSFKMqsZcItlg.gif" /></figure><h3>Why Data Profiling Is Essential</h3><p>Data Profiling will help you to answer questions like:</p><ul><li>Is my data complete?</li><li>Are columns the right data type?</li><li>Are values consistently formatted (case, spacing)?</li><li>Are there outliers that might distort my analysis?</li><li>Are there duplicates?</li><li>How many unique values do I have per column?</li></ul><p>You can’t build meaningful dashboards and reliable ML models without answering all of the above questions first.</p><h3>Building a Data Profiling Dashboard in Power BI with Python</h3><h4>About the Dataset</h4><p>I used the <a href="https://www.kaggle.com/datasets/ahmedmohamed2003/cafe-sales-dirty-data-for-cleaning-training/data"><strong>Cafe Sales</strong></a> dataset (intentionally messy data from Kaggle, I made it even messier for our example) and did some pre-cleaning (standardized various error tokens like “error”, “ ”, “Nan”→ NaN) so the profiling script can run consistently.</p><p>In the end of this article you will find the whole code ready to use it but for now, let’s explain it step by step to help you understand the logic behind it.</p><h4>1. Basic Overview</h4><p>I wanted first to have a quick way to check for missing data, incorrect data types or unexpected uniqueness at a glance. So, I started by building an overview table.</p><pre># --- import required libraries --- <br>import pandas as pd <br>import numpy as np<br><br># --- 1. BASIC OVERVIEW --- <br>profile = pd.DataFrame({ <br>    &quot;Column&quot;: dataset.columns, <br>    &quot;Data_Type&quot;: dataset.dtypes.astype(str), <br>    &quot;Missing_Values&quot;: dataset.isnull().sum(), <br>    &quot;Missing_%&quot;: (dataset.isnull().mean() * 100).round(2), <br>    &quot;Unique_Values&quot;: dataset.nunique(), <br>    &quot;Total_Values&quot;: len(dataset)<br>})</pre><p>What it does:</p><p>It builds a profile DataFrame that summarizes each column in our dataset.</p><ul><li><strong>Column </strong>: column name</li><li><strong>Data_Type </strong>: pandas dtype as a string (e.g. int64, object, datetime)</li><li><strong>Missing_Values </strong>: number of NaN in that column</li><li><strong>Missing_% </strong>: percentage of missing values (rounded to 2 decimals)</li><li><strong>Uniques_Values </strong>: count of unique values in the column</li><li><strong>Total_Values </strong>: total rows in dataset (same number repeated for every row)</li></ul><p>Outcome:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/769/1*qetVsHK8j17bW71jr7hpGA.png" /></figure><h4>2. TEXT CONSISTENCY CHECK</h4><p>Text columns often hide data quality problems like inconsistent casing, extra spaces, mixed formats (upper case or lower case errors). If you haven’t properly cleaned them you will definitely face problems with joins, grouping or filters. So, it is important to have a table that will check about text problems.</p><pre># Select object-type columns with at least 1 non-null value<br>text_cols = [<br>    col for col in dataset.select_dtypes(include=[&#39;object&#39;]).columns<br>    if dataset[col].notna().sum() &gt; 0<br>]</pre><p>What it does:</p><ul><li>Selects object type columns but only those that have at least one not null value (to avoid wasting compute on empty text columns).</li></ul><pre><br>text_report = []<br><br>for col in text_cols:<br>    sample = dataset[col].dropna().astype(str)<br>    total_rows = len(sample)<br>    if total_rows == 0:<br>        continue<br>    <br>    # Count casing patterns<br>    all_upper = (sample == sample.str.upper()).sum()<br>    all_lower = (sample == sample.str.lower()).sum()<br>    title_case = (sample == sample.str.title()).sum()  <br><br>    # Detect leading/trailing spaces <br>    spaces = sample[sample.str.strip() != sample].count()<br>       <br>    # Calculate uniformity (percentage of dominant style)<br>    dominant_case = max(all_upper, all_lower, title_case)<br>    case_uniformity = round((dominant_case / total_rows) * 100, 2)<br><br>    # Additional consistency checks<br>    numeric_like = sample.str.isnumeric().sum()<br>    contains_digits = sample.str.contains(r&quot;\d&quot;).sum()<br><br>    text_report.append({<br>        &quot;Column&quot;: col,<br>        &quot;Total_Values&quot;: total_rows,<br>        &quot;Unique_Values&quot;: sample.nunique(),<br>        &quot;All_Uppercase_Count&quot;: int(all_upper),<br>        &quot;All_Lowercase_Count&quot;: int(all_lower),<br>        &quot;Title_Case_Count&quot;: int(title_case),<br>        &quot;Leading_Trailing_Spaces&quot;: int(spaces),<br>        &quot;Numeric_Like_Count&quot;: int(numeric_like),<br>        &quot;Contains_Digits_Count&quot;: int(contains_digits),<br>        &quot;Case_Uniformity_%&quot;: case_uniformity<br>    })<br><br>text_profile = pd.DataFrame(text_report)</pre><p>What it does:</p><ul><li><strong>sample : </strong>Non null values as strings</li><li><strong>all_upper : </strong>count of values that are fully uppercase</li><li><strong>all_lower : </strong>count of values that are fully lowercase</li><li><strong>title_case : </strong>count of values where first letter capitalized each word</li><li><strong>spaces : </strong>count of values that have leading/trailing spaces</li><li><strong>dominant case : </strong>most common casing pattern</li><li><strong>case_uniformity : </strong>percent of the column that matches that dominant pattern</li><li><strong>numeric_like : </strong>count of values that are only digits (useful if numeric values were imported as text)</li><li><strong>contains_digits : </strong>count of text values that contain any digit</li></ul><p>Outcome:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/368/1*RRJ67YIqgxtwulQ15vKGkg.png" /></figure><p>Tip: I unpivoted the columns of this table for visual purposes.</p><h4>3. OUTLIER DETECTION</h4><p>One more think that would help us to observe the quality of our data is to check for outliers. This way we can quickly surface suspicious values (wrong units, bad parsing, real extreme values) and decide whether to keep, transform, or investigate further those values.</p><p>Tip: I used the IQR method here to detect outliers.</p><pre>numeric_cols = dataset.select_dtypes(include=np.number).columns<br><br>outlier_report = []<br><br>for col in numeric_cols:<br>    q1 = dataset[col].quantile(0.25)<br>    q3 = dataset[col].quantile(0.75)<br>    iqr = q3 - q1<br>    lower_bound = q1 - 1.5 * iqr<br>    upper_bound = q3 + 1.5 * iqr<br>    outliers = dataset[(dataset[col] &lt; lower_bound) | (dataset[col] &gt; upper_bound)][col]<br>   <br>    outlier_report.append({<br>        &quot;Column&quot;: col,<br>        &quot;Min&quot;: dataset[col].min(),<br>        &quot;Max&quot;: dataset[col].max(),<br>        &quot;Lower_Bound&quot;: round(lower_bound, 2),<br>        &quot;Upper_Bound&quot;: round(upper_bound, 2),<br>        &quot;Outlier_Count&quot;: len(outliers),<br>        &quot;Outlier_%&quot;: round(len(outliers) / len(dataset) * 100, 2)<br>    })<br><br>outlier_profile = pd.DataFrame(outlier_report)</pre><p>What it does:</p><ul><li><strong>numeric_cols :</strong> select all the numerical type columns</li><li>calculates <strong>Q1, Q3, IQR</strong></li><li>computes standard Tukey lower/upper bounds (<strong>Q1 — 1.5*IQR, Q3 + 1.5*IQR</strong>)</li><li>counts the values outside those bounds as outliers</li><li>returns min/max, lower and upper bounds, the outlier counts and percentage.</li></ul><p>Outcome:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/285/1*tHoFpVe9PPtHQfRd0uA9tA.png" /></figure><p>Tip: I unpivoted the columns of this table for visual purposes.</p><h4>4. DUPLICATE CHECK</h4><p>Duplicated rows can change total, averages and model training. So, it is a important aspect of our quality check.</p><pre># Check for full row duplicates<br>total_duplicates = dataset.duplicated().sum()<br><br>duplicate_rows = dataset[dataset.duplicated(keep=False)] <br><br>duplicates_summary = pd.DataFrame({<br>    &quot;Total_Rows&quot;: [len(dataset)],<br>    &quot;Duplicate_Rows_Count&quot;: [len(duplicate_rows)],<br>    &quot;Unique_Rows_Count&quot;: [len(dataset) - total_duplicates],<br>    &quot;Duplicates_%&quot;: [round((total_duplicates / len(dataset)) * 100, 2)]<br>})</pre><p>What it does:</p><ul><li><strong>dataset.duplicated() </strong>: marks rows that are duplicates of a previous row.</li><li><strong>total_duplicates : </strong>counts them.</li><li><strong>duplicate_rows :</strong> extracts all rows that appear duplicated (keeps all occurrences).</li><li><strong>duplicates_summary : </strong>creates a summary table with counts and percent duplicates.</li></ul><p>Outcomes:</p><p>A table with all the duplicated rows:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/980/1*lV1amoqMxcnZuYsbJgo7NA.png" /></figure><p>A summary table:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/638/1*4zBH_CEMqOqaKKGGw6Z0EA.png" /></figure><h4>5. Accuracy Rules (custom rules)</h4><p>I also wanted to create custom rules to check how many values pass each rule to observe the accuracy of our dataset. In this way, we can check not only generic profiling like missing data and data types but to also <strong>catch logically incorrect values.</strong></p><pre>accuracy_rules = []<br><br># Rule 1: Price Per Unit &gt; 0<br>price_valid = (dataset[&#39;Price Per Unit&#39;] &gt; 0).sum()<br>price_total = dataset[&#39;Price Per Unit&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;:&#39;Price Per Unit&#39;,<br>    &#39;Valid_Count&#39;: price_valid,<br>    &#39;Total_Count&#39;: price_total,<br>    &#39;Accuracy_%&#39;: round(price_valid / price_total * 100, 2)<br>})<br><br># Rule 2: Total Spent &gt; 0<br>Total_Spent_valid = (dataset[&#39;Total Spent&#39;] &gt; 0).sum()<br>Total_Spent_total = dataset[&#39;Total Spent&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;: &#39;Total Spent&#39;,<br>    &#39;Valid_Count&#39;: Total_Spent_valid,<br>    &#39;Total_Count&#39;: Total_Spent_total,<br>    &#39;Accuracy_%&#39;: round(Total_Spent_valid / Total_Spent_total * 100, 2)<br>})<br><br># Rule 3: Quantity between 1 and 100<br>qty_valid = ((dataset[&#39;Quantity&#39;] &gt;= 1) &amp; (dataset[&#39;Quantity&#39;] &lt;= 100)).sum()<br>qty_total = dataset[&#39;Quantity&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;: &#39;Quantity&#39;,<br>    &#39;Valid_Count&#39;: qty_valid,<br>    &#39;Total_Count&#39;: qty_total,<br>    &#39;Accuracy_%&#39;: round(qty_valid / qty_total * 100, 2)<br>})<br><br>accuracy = pd.DataFrame(accuracy_rules)</pre><p>What it does:</p><p>Checks if:</p><ul><li>Price Per Unit &gt; 0,</li><li>Total Spent &gt; 0,</li><li>Quantity is between 1 and 100.</li></ul><p>Outcome:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/469/1*HPGlevIFINrdpb1L_WLW0Q.png" /></figure><h4><strong>6. SUMMARY STATISTICS </strong>(numeric &amp; object)</h4><p>With the statistics summary we have a quick distribution and central tendency info for both numeric and categorical fields.</p><pre>summary_stats_obj = dataset.describe(include=&#39;object&#39;).T.reset_index().rename(columns={&#39;index&#39;: &#39;Column&#39;})<br>summary_stats_num = dataset.describe(include=np.number).T.reset_index().rename(columns={&#39;index&#39;: &#39;Column&#39;})</pre><p>What it does</p><ul><li><strong>describe(include=object)</strong> : returns summary stats for text columns: count, unique, top, freq etc.</li><li><strong>describe(include=np.number)</strong> : returns numeric stats: count, mean, std, min, 25%, 50%, 75%, max.</li><li>The <strong>T.reset_index().rename(...)</strong> : converts the result so each row corresponds to a column (makes it easy to display in Power BI).</li></ul><p>Outcome:</p><p>Statistics table for numeric type columns</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/281/1*CU7Y3zahLv9hlP22TVmVEg.png" /></figure><p>Statistics Table for object type columns:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/324/1*-YGCjrUDwst0dEeXArzjMA.png" /></figure><h4>7. OUTPUTS</h4><pre># You can duplicate and rename in Power BI to show as separate tables<br>profile, text_profile, outlier_profile, duplicate_rows, duplicates_summary, accuracy, summary_stats_obj, summary_stats_num</pre><p>What it does</p><ul><li>Power BI captures any DataFrames returned by the script and turns them into query outputs (you will see them as separate tables in Power Query).</li><li>The order you return them is the order Power BI shows them — you can duplicate/rename in Power Query to create friendly table names like Profile, TextProfile, Outliers etc.</li></ul><h4>Final Code</h4><pre># --- Data Profiling Script for Power BI ---<br><br># --- Import Required Libraries ---<br>import pandas as pd<br>import numpy as np<br><br># --- 1. BASIC OVERVIEW ---<br><br>profile = pd.DataFrame({<br>    &quot;Column&quot;: dataset.columns,<br>    &quot;Data_Type&quot;: dataset.dtypes.astype(str),<br>    &quot;Missing_Values&quot;: dataset.isnull().sum(),<br>    &quot;Missing_%&quot;: (dataset.isnull().mean() * 100).round(2),<br>    &quot;Unique_Values&quot;: dataset.nunique(),<br>    &quot;Total_Values&quot;: len(dataset),<br>    &quot;Total_Rows&quot;: len(dataset)<br>})<br><br># --- 2. TEXT CONSISTENCY CHECK (for object columns) ---<br><br># Select object-type columns with at least 1 non-null value<br>text_cols = [<br>    col for col in dataset.select_dtypes(include=[&#39;object&#39;]).columns<br>    if dataset[col].notna().sum() &gt; 0<br>]<br><br>text_report = []<br><br>for col in text_cols:<br>    sample = dataset[col].dropna().astype(str)<br>    total_rows = len(sample)<br>    if total_rows == 0:<br>        continue<br><br>    # Count casing patterns<br>    all_upper = (sample == sample.str.upper()).sum()<br>    all_lower = (sample == sample.str.lower()).sum()<br>    title_case = (sample == sample.str.title()).sum()  <br><br>    # Detect leading/trailing spaces<br>    spaces = sample[sample.str.strip() != sample].count()<br><br>    # Calculate uniformity (percentage of dominant style)<br>    dominant_case = max(all_upper, all_lower, title_case)<br>    case_uniformity = round((dominant_case / total_rows) * 100, 2)<br><br>    # Additional consistency checks<br>    numeric_like = sample.str.isnumeric().sum()<br>    contains_digits = sample.str.contains(r&quot;\d&quot;).sum()<br><br>    text_report.append({<br>        &quot;Column&quot;: col,<br>        &quot;Total_Values&quot;: total_rows,<br>        &quot;Unique_Values&quot;: sample.nunique(),<br>        &quot;All_Uppercase_Count&quot;: int(all_upper),<br>        &quot;All_Lowercase_Count&quot;: int(all_lower),<br>        &quot;Title_Case_Count&quot;: int(title_case),<br>        &quot;Leading_Trailing_Spaces&quot;: int(spaces),<br>        &quot;Numeric_Like_Count&quot;: int(numeric_like),<br>        &quot;Contains_Digits_Count&quot;: int(contains_digits),<br>        &quot;Case_Uniformity_%&quot;: case_uniformity<br>    })<br><br>text_profile = pd.DataFrame(text_report)<br><br># --- 3. OUTLIER DETECTION (IQR method) ---<br><br>numeric_cols = dataset.select_dtypes(include=np.number).columns<br><br>outlier_report = []<br><br>for col in numeric_cols:<br>    q1 = dataset[col].quantile(0.25)<br>    q3 = dataset[col].quantile(0.75)<br>    iqr = q3 - q1<br>    lower_bound = q1 - 1.5 * iqr<br>    upper_bound = q3 + 1.5 * iqr<br><br>    outliers = dataset[(dataset[col] &lt; lower_bound) | (dataset[col] &gt; upper_bound)][col]<br><br>    outlier_report.append({<br>        &quot;Column&quot;: col,<br>        &quot;Min&quot;: dataset[col].min(),<br>        &quot;Max&quot;: dataset[col].max(),<br>        &quot;Lower_Bound&quot;: round(lower_bound, 2),<br>        &quot;Upper_Bound&quot;: round(upper_bound, 2),<br>        &quot;Outlier_Count&quot;: len(outliers),<br>        &quot;Outlier_%&quot;: round(len(outliers) / len(dataset) * 100, 2)<br>    })<br><br>outlier_profile = pd.DataFrame(outlier_report)<br><br># --- 4. DUPLICATE CHECK ---<br><br># Check for full row duplicates<br>total_duplicates = dataset.duplicated().sum()<br><br>duplicate_rows = dataset[dataset.duplicated(keep=False)] <br><br>duplicates_summary = pd.DataFrame({<br>    &quot;Total_Rows&quot;: [len(dataset)],<br>    &quot;Duplicate_Rows_Count&quot;: [len(duplicate_rows)],<br>    &quot;Unique_Rows_Count&quot;: [len(dataset) - total_duplicates],<br>    &quot;Duplicates_%&quot;: [round((total_duplicates / len(dataset)) * 100, 2)]<br>})<br><br><br># --- 5. ACCURACY RULES ---<br><br>accuracy_rules = []<br><br># Rule 1: Price Per Unit &gt; 0<br>price_valid = (dataset[&#39;Price Per Unit&#39;] &gt; 0).sum()<br>price_total = dataset[&#39;Price Per Unit&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;:&#39;Price Per Unit&#39;,<br>    &#39;Valid_Count&#39;: price_valid,<br>    &#39;Total_Count&#39;: price_total,<br>    &#39;Accuracy_%&#39;: round(price_valid / price_total * 100, 2)<br>})<br><br># Rule 2: Total Spent &gt; 0<br>Total_Spent_valid = (dataset[&#39;Total Spent&#39;] &gt; 0).sum()<br>Total_Spent_total = dataset[&#39;Total Spent&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;: &#39;Total Spent&#39;,<br>    &#39;Valid_Count&#39;: Total_Spent_valid,<br>    &#39;Total_Count&#39;: Total_Spent_total,<br>    &#39;Accuracy_%&#39;: round(Total_Spent_valid / Total_Spent_total * 100, 2)<br>})<br><br># Rule 3: Quantity between 1 and 100<br>qty_valid = ((dataset[&#39;Quantity&#39;] &gt;= 1) &amp; (dataset[&#39;Quantity&#39;] &lt;= 100)).sum()<br>qty_total = dataset[&#39;Quantity&#39;].notnull().sum()<br>accuracy_rules.append({<br>    &#39;Column&#39;: &#39;Quantity&#39;,<br>    &#39;Valid_Count&#39;: qty_valid,<br>    &#39;Total_Count&#39;: qty_total,<br>    &#39;Accuracy_%&#39;: round(qty_valid / qty_total * 100, 2)<br>})<br><br>accuracy = pd.DataFrame(accuracy_rules)<br><br><br># --- 6. SUMMARY STATISTICS (numeric &amp; object)---<br><br>summary_stats_obj = dataset.describe(include=&#39;object&#39;).T.reset_index().rename(columns={&#39;index&#39;: &#39;Column&#39;})<br><br>summary_stats_num = dataset.describe(include=np.number).T.reset_index().rename(columns={&#39;index&#39;: &#39;Column&#39;})<br><br><br># --- 7. OUTPUT TABLES FOR POWER BI ---<br># You can duplicate and rename in Power BI to show as separate tables<br>profile, text_profile, outlier_profile, duplicate_rows, duplicates_summary, accuracy, summary_stats_obj, summary_stats_num</pre><h3>Final Thoughts</h3><p>This script creates a compact profiling pipeline inside Power BI using Python. It:</p><ol><li>produces column-level metadata (types, missing, uniques),</li><li>runs deep text consistency checks (case, spaces, numeric-like),</li><li>detects numeric outliers using IQR,</li><li>surfaces duplicate rows and a small duplicates summary,</li><li>evaluates business accuracy rules (e.g., positive prices),</li><li>outputs both object &amp; numeric summary statistics.</li></ol><p>Return these DataFrames to Power BI as separate query tables and wire them into interactive visuals to accelerate your data-cleaning decisions.</p><h3>What’s next (Part 2)</h3><p>In Part 2 I’ll show you how to:</p><ul><li>turn these outputs into a clean profiling dashboard in Power BI (layout, conditional formatting, bookmarks to switch context),</li><li>wire filters and a column selector so you can inspect any column interactively.</li></ul><p>🌟 If you enjoyed this article dont’ forget to 👏 clap and follow for more related content!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6508a43df551" width="1" height="1" alt=""><hr><p><a href="https://medium.com/codex/how-to-detect-hidden-data-issues-in-power-bi-with-a-python-powered-profiling-dashboard-part-1-6508a43df551">How to Detect Hidden Data Issues in Power BI with a Python-Powered Profiling Dashboard (Part 1)</a> was originally published in <a href="https://medium.com/codex">CodeX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>