<?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 Ari Sulistiyo Prabowo on Medium]]></title>
        <description><![CDATA[Stories by Ari Sulistiyo Prabowo on Medium]]></description>
        <link>https://medium.com/@dataimpact?source=rss-125434829b03------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*FKxUp3l8dXxIOI4q6go8SA.jpeg</url>
            <title>Stories by Ari Sulistiyo Prabowo on Medium</title>
            <link>https://medium.com/@dataimpact?source=rss-125434829b03------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 02:39:52 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@dataimpact/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[From PDF to RAG: Building an AI Assistant with Python and Hugging Face]]></title>
            <link>https://medium.com/cultigo/from-pdf-to-rag-building-an-ai-assistant-with-python-and-hugging-face-cf631ea0eaf4?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/cf631ea0eaf4</guid>
            <category><![CDATA[retrieval-augmented-gen]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Thu, 31 Jul 2025 01:31:51 GMT</pubDate>
            <atom:updated>2025-07-31T02:17:56.004Z</atom:updated>
            <content:encoded><![CDATA[<p>Have you ever wished you could ask a long, complicated PDF document a quick question and get a direct answer? Imagine this: you upload your company’s user manual, your car manual document, and internal department documentation, and then ask:</p><blockquote>“Could you explain how to activate Adaptive Cruise Control in my car?”<br>or<br>“How to claim company’s insurace when I get sick and who I should call?”</blockquote><p>Instead of skimming through several pages of text, an AI assistant instantly gives you the right answer, straight from your document.</p><p>That’s what we call Retrieval-Augmented Generation (RAG), a smart AI that can search your documents and answer questions with clarity and context.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a4ta-u2KCwfinYdbWyS94w.png" /><figcaption>AI Assistant for Smooth Onboarding of New Joiners</figcaption></figure><h3>What is RAG, and why is it powerful?</h3><p>RAG stands for Retrieval-Augmented Generation. Let’s break it down:</p><ul><li><strong>Retriever:</strong> Looks through your documents to find the most relevant parts.</li><li><strong>Generator:</strong> Uses an LLM to craft a natural-sounding answer based on what it finds.</li></ul><p>Here is the analogy: RAG is like a smart librarian, and imagine walking into a library and asking, “What is the terminology of Net Sales in this business unit?”</p><ul><li>You need to explore by searching for the keywords “Net Sales” and spend several minutes searching and validating.</li><li>But a RAG AI acts like a librarian. It first finds the most relevant pages in your company documentation (retrieval), then explains the answer in plain language (generation).</li></ul><h3>Step 1: Extracting Knowledge from PDFs</h3><p>Most companies store knowledge in documentation platforms such as Google Docs, PDF, and Notion.</p><p>To teach our AI assistant, we first need to extract text from the PDF. Here’s how to do it using Python, DirectoryLoader, and PyPDF library:</p><pre>from langchain_community.document_loaders import DirectoryLoader, PyPDFLoader<br><br># List folders<br>folders = glob.glob(&quot;knowledge/*&quot;)<br><br>documents = []<br><br>for folder in folders:<br>    doc_type = os.path.basename(folder)<br><br>    # Replace TextLoader with PDF-specific loader<br>    loader = DirectoryLoader(<br>        folder,<br>        glob=&quot;**/*.pdf&quot;,<br>        loader_cls=PyPDFLoader  <br>    )<br><br>    folder_docs = loader.load()<br>    for doc in folder_docs:<br>        doc.metadata[&quot;doc_type&quot;] = doc_type<br>        documents.append(doc)</pre><p>This function will load your folder, find everything in PDF format, read each page, and return the full text as a string.</p><h3>Step 2: Chunking the Text (AI Can Read it)</h3><p>The reason behind chunking the text is Large Language Model (like GPT) can only read a limited number of words (called “tokens”) at a time. If your PDF document is 50 pages, it’s too much.</p><p>So we break the document into smaller overlapping pieces called chunks.</p><pre>from langchain.text_splitter import CharacterTextSplitter<br><br># Splitting text into smaller chunks<br>text_splitter = CharacterTextSplitter(chunk_size=500, chunk_overlap=50)<br>chunks = text_splitter.split_documents(documents)<br><br>print(f&quot;Total chunks created: {len(chunks)}&quot;)</pre><p>Think of it like breaking a long article into smaller paragraphs, so the AI can read and understand them one at a time.</p><h3>Step 3: Making Text Searchable with Embeddings</h3><p>Now we need to make these chunks searchable. For that, we use embeddings, a fancy term for turning text into numbers so the AI can “compare meanings”.</p><blockquote>Think of embeddings as numerical fingerprint of text</blockquote><p>When you pass a sentence through an embedding model (like OpenAI Embeddings), it converts that sentence into a list of numbers called a vector. These numbers capture the semantic meaning (how words get their meaning) of the text.</p><p>Here is how to do it with OpenAI Embeddings:</p><pre>from langchain_openai import OpenAIEmbeddings<br><br>embeddings = OpenAIEmbeddings()</pre><p>Each chunk is now a vector, a list of numbers that represents the meaning of that text. We store these vectors in a vector database like FAISS.</p><blockquote><strong>Why do we need a vector database?</strong></blockquote><p>We don’t just have 5 or 10 chunks of text, but we will have hundreds or thousands of chunks. We want to:</p><ul><li>Convert the question into a vector (embedding)</li><li>Find the most similar vectors among all the document chunks (based on meaning)</li></ul><p>This is called vector similarity search, and it’s handled efficiently by a vector database.</p><pre>from langchain_community.vectorstores import FAISS<br><br># Create vectorstore<br>vectorstore = FAISS.from_documents(chunks, embedding=embeddings)</pre><h3>Step 4: Building the RAG Pipeline</h3><p>It’s time to connect everything:</p><ol><li>Users ask a question</li><li>AI retrieves the top relevant chunks from your documents</li><li>AI generates an answer using the LLM</li></ol><p>Here is a simplified example using LangChain:</p><pre>from langchain_core.prompts import ChatPromptTemplate<br>from langchain.memory import ConversationBufferMemory<br>from langchain.chains.combine_documents import create_stuff_documents_chain<br>from langchain.chains import create_retrieval_chain<br>from langchain_openai import ChatOpenAI<br><br># Your vectorstore and retriever setup<br>retriever = vectorstore.as_retriever()<br><br>prompt = ChatPromptTemplate.from_messages([<br>    (&quot;system&quot;, &quot;You are a helpful assistant.&quot;),<br>    (&quot;human&quot;, &quot;{chat_history}\n\nContext:\n{context}\n\nUser: {input}&quot;)<br>])<br><br><br># 2. Set up memory<br>memory = ConversationBufferMemory(<br>    memory_key=&quot;chat_history&quot;,<br>    input_key=&quot;input&quot;,<br>    output_key=&quot;output&quot;,<br>    return_messages=True<br>)<br><br># 3. Set up document chain<br>llm = ChatOpenAI(model=MODEL, temperature=0)<br>document_chain = create_stuff_documents_chain(llm, prompt)<br><br># 4. Retrieval chain<br>retrieval_chain = create_retrieval_chain(retriever, document_chain)<br><br># 5. Chat loop example (or single turn)<br>query = &quot;Could you provide the reference for payment?&quot;<br>history_str = memory.buffer_as_str  # Get formatted chat history<br><br>result = retrieval_chain.invoke({<br>    &quot;input&quot;: query,<br>    &quot;chat_history&quot;: history_str,<br>})<br><br># Update memory with the new conversation<br>memory.save_context({&quot;input&quot;: query}, {&quot;output&quot;: result[&quot;answer&quot;]})<br><br>print(result[&quot;answer&quot;])</pre><p>The question is matched to relevant chunks using vector search, then passed to GPT to generate a natural answer.</p><h3>Step 5: Deploy to Hugging Face Spaces (for Free!)</h3><p>I chose Streamlit as the UI interface for my AI assistant and deployed it to Hugging Face Spaces. Here is the flowchart for deployment:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/807/1*MG5sknz1m3geSEcxdn62RA.png" /><figcaption>Flowchart RAG AI Deployment Process</figcaption></figure><p>In Hugging Face Spaces, you need to prepare several files:</p><ul><li>requirements.txt → to install several Python libraries for your app.</li><li>faiss index → the saved vector database from your local notebook.</li><li>RAG pipeline (app.py) → the framework of your RAG AI.</li><li>UI (streamlit.py) → UI framework.</li></ul><p>You can refer to my <a href="https://github.com/densaiko/AI-Project/tree/main/RAG%20AI%20Deployment">GitHub Repositories</a> for detailed code.</p><h3>Bonus: Tips to Make It Smarter</h3><ul><li>Use <strong>ConversationalRetrievalChain</strong> for memory (chat-like responses)</li><li>You can also swap OpenAI with open-source models like Mistral or others to see the best performance for your AI assistant.</li></ul><h3>Conclusion</h3><p>RAG AI is not new when I write this article, and many people and companies have implemented this AI technology to help their internal parties and even their customers/clients.</p><p>The above step-by-step process is designed to simplify the learning experience for you. Now it’s time to start small with the simple one, then build based on the challenge that you want to solve using RAG AI.</p><p>Feel free to connect with me on <a href="https://www.linkedin.com/in/ariprabowo/">LinkedIn</a>. I would be delighted to share my experiences in the tech industry, data science, marketing, and product. Thank you for reading my article.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cf631ea0eaf4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/cultigo/from-pdf-to-rag-building-an-ai-assistant-with-python-and-hugging-face-cf631ea0eaf4">From PDF to RAG: Building an AI Assistant with Python and Hugging Face</a> was originally published in <a href="https://medium.com/cultigo">Cultigo</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Gemini AI Supercharged My BigQuery SQL Workflow]]></title>
            <link>https://medium.com/cultigo/how-gemini-ai-supercharged-my-bigquery-sql-workflow-c83240760574?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/c83240760574</guid>
            <category><![CDATA[bigquery]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[sql]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[gemini]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Sun, 01 Jun 2025 08:45:33 GMT</pubDate>
            <atom:updated>2025-06-01T08:45:33.649Z</atom:updated>
            <content:encoded><![CDATA[<p>Starting a new job often comes with a common challenge: figuring out how to understand and navigate the data in your company’s data warehouse, especially if you’re using a platform like BigQuery. In the first month, it’s normal to feel unsure about where to begin.</p><p>To make things more complicated, you’ll likely encounter dozens of tables that look similar. Your first task might be as simple as querying the data, but you quickly find yourself spending 60% of your time cleaning data and designing queries, instead of focusing on the actual analysis.</p><blockquote>In the age of AI-assisted coding, the ability to extract narrative and impact from data is what makes you irreplaceable.</blockquote><p>With the overwhelming amount of data, it’s easy to feel stuck and waste valuable time figuring out where to start. But that’s where <strong>Gemini AI in BigQuery</strong> comes in — it’s designed to boost your productivity by helping you explore, query, and analyze data more efficiently.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kRsI063VkgBwm5fFTvbGhw.png" /><figcaption>Image is generated by ChatGPT</figcaption></figure><h4>Data Exploration with Gemini in BigQuery</h4><p>The first step you should consider is exploring a single table. BigQuery provides a helpful feature called <strong>Table Explorer</strong>, which shows the unique values in each column. It’s a great starting point to understand the data structure without having to scroll through endless rows or write a DISTINCT query just to check for uniqueness.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rdBGT9QD6SzMdKGZpGwG8g.png" /><figcaption>Table Explorer shows distinct values instantly without SQL (Screenshot from BigQuery Studio)</figcaption></figure><p>Sometimes your manager or stakeholders may ask for quick insights from the table you’ve just explored — or maybe you’re just curious yourself. In that case, you can use the Insights feature, which offers a list of AI-generated suggestions and trends that you can further explore.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uUcg-Bn9ewkx1jHS8Hyw1g.png" /><figcaption>BigQuery’s Insight feature offers ready-to-explore data questions (Screenshot from BigQuery Studio)</figcaption></figure><blockquote>Always start by working with a single table. Exploring one table thoroughly is far more efficient than immediately jumping into complex multi-table joins.</blockquote><h4>Developing Code with Gemini in BigQuery</h4><p>If you’re a data analyst or data scientist, this is likely where you’ll spend most of your time — not cleaning data, but querying it efficiently to build reports, perform analysis, or prepare data for machine learning development.</p><p>In BigQuery, Gemini AI supports your workflow in two key ways:</p><ul><li>SQL Generation</li><li>SQL Explanation</li></ul><p><strong>SQL Generation, t</strong>his feature focuses on generating queries to help you achieve your desired output. Gemini offers two methods for generating SQL:</p><ul><li>Using a hashtag (#) to prompt directly in the query editor</li><li>Using the pencil icon in the input field to enter a natural language prompt</li></ul><p>Both approaches allow you to write in plain English. Just describe what you need, and Gemini will generate the SQL for you. Start simple — a clear prompt usually gets the best result.</p><blockquote># How many claims in each product that the claim status is paid, show the name of the product.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/902/1*1YlvYcbaZIAoEHnYIix69w.png" /><figcaption>Prompting Gemini AI directly in BigQuery to generate SQL based on natural language input (Screenshot from BigQuery Studio)</figcaption></figure><p>You can use the <strong>same prompt</strong> in either method — via hashtag or the pencil icon on the left side of the query box.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/917/1*EsRLzDoVlkcXJnQHKmS__w.png" /><figcaption>Prompting Gemini AI using the pencil feature (Screenshot from BigQuery Studio)</figcaption></figure><p>On the <strong>SQL Explanation</strong>, this feature helps you understand the logic behind a query , especially useful when Gemini introduces a function you may not be familiar with. It explains how the query was structured and how different parts contribute to the final result.</p><p>In short, SQL Explanation gives you a step-by-step breakdown, helping you follow the logic from start to finish.</p><h4>What If the Query from Gemini Is Incorrect?</h4><p>Sometimes, the generated query may not meet your expectations for example, it might apply the wrong join between tables, include incorrect GROUP BY clauses, or select misleading columns. In cases like this, you can use the <strong>Transform </strong>feature in Gemini.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/783/1*eauqMaJUb-GcQEQHnKEI0w.png" /><figcaption>Using Gemini’s Transform feature to revise and improve an initial SQL query (Screenshot from BigQuery Studio)</figcaption></figure><p>To transform a query:</p><ol><li>Highlight the part of the query you want to fix</li><li>Click the Gemini logo on the left side of the query window</li><li>Select Transform</li></ol><p>Gemini will suggest improvements or corrections based on your selected query segment.</p><h4>Visualizing Data and Designing Queries in Data Canvas</h4><p>One of the challenges when using Gemini for query generation is the potential for errors when joining tables, often caused by prompts that are either too vague or overly specific.</p><p>That’s where <strong>Data Canvas</strong> comes in. It’s designed to feel like working on a canvas, allowing you to visually select tables, generate queries, create visualizations, and extract insights all in one interactive flow.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*11Zi3_NaH18XqanwYfd1zg.png" /><figcaption>Data Canvas in BigQuery (Screenshot from BigQuery Studio)</figcaption></figure><p>Once you’ve selected the tables, you can begin prompting Gemini to join them. I recommend joining tables only using the columns that will help you derive meaningful insights , this keeps the results clean and relevant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/948/1*pfVYpLYfliZjR7Cy19kUBA.png" /><figcaption>Prompting Gemini to join selected tables in BigQuery (Screenshot from BigQuery Studio)</figcaption></figure><p>After the tables are joined and the relevant columns selected, you have the option to <strong>“Query these results”</strong>. From there, you can prompt Gemini to generate aggregated data, such as calculating the total insurance claims per region.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-rGBl448dquatCXOP62QAA.png" /><figcaption>Querying the joined data in Data Canvas to get aggregated claim results by region (Screenshot from BigQuery Studio)</figcaption></figure><p>When you’ve obtained aggregated results, you can use the <strong>Visualize</strong> option to display your data as bar charts, line charts, pie charts, or even custom visuals. You can refine your charts by sorting, changing colors, or customizing them further using natural language prompts.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZN3ePtCzf54KqQs7f4V99A.png" /><figcaption>BigQuery’s visualization tool displays regional claim counts clearly (Screenshot from BigQuery Studio)</figcaption></figure><p>Finally, the <strong>Generate Insights</strong> feature provides a summary of your results, highlighting key patterns such as top-performing or least-performing products. It’s a great finishing touch for quickly grasping trends and outliers in your data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/990/1*4RnjE7Q3AsuxnfPBTr_Tvg.png" /><figcaption>Generate Insights provides quick takeaways from your aggregated data (Screenshot from BigQuery Studio)</figcaption></figure><h4>Conclusion</h4><p>Gemini AI in BigQuery is incredibly helpful, not just for new team members who are trying to understand the data in their first month, but also for existing data professionals who want to boost productivity. It allows your team to focus more on analysis and insight generation, rather than spending excessive time writing queries for every new task.</p><p>These features aren’t new, but if your BigQuery environment hasn’t adopted them yet, now is a great time to start. Talk with your team and explore how Gemini AI can help you deliver results faster and more efficiently.</p><p>Feel free to connect with me on <a href="https://www.linkedin.com/in/ariprabowo/">LinkedIn</a>. I would be delighted to share my experiences in the tech industry, data science, marketing, and product.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c83240760574" width="1" height="1" alt=""><hr><p><a href="https://medium.com/cultigo/how-gemini-ai-supercharged-my-bigquery-sql-workflow-c83240760574">How Gemini AI Supercharged My BigQuery SQL Workflow</a> was originally published in <a href="https://medium.com/cultigo">Cultigo</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Boosting Tableau Performance: Strategies for Faster and More Efficient Dashboards]]></title>
            <link>https://medium.com/cultigo/boosting-tableau-performance-strategies-for-faster-and-more-efficient-dashboards-2a4eb171e737?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/2a4eb171e737</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[decision-making]]></category>
            <category><![CDATA[tableau]]></category>
            <category><![CDATA[performance]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Fri, 26 Jul 2024 01:40:32 GMT</pubDate>
            <atom:updated>2024-07-26T01:40:32.077Z</atom:updated>
            <content:encoded><![CDATA[<p>Facing low speed in loading dashboards in Business Intelligence tools can impact users who consume the data and make decisions from the dashboard. As a result, users feel unproductive while waiting for the dashboard to load. Do you face the same issue in your company today? If so, it would be best to revamp your dashboard based on what I will share below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LZO8BwcgHVzAcnJU3UQJWQ.png" /><figcaption>Photo from Author</figcaption></figure><h3>Why Speed and Performance are Important?</h3><p>Tableau workbook performance drives speed to insight, analytical flow, and iterative development. Optimizing dashboard performance is one of the most important responsibilities for an analyst. Let’s break this down further:</p><ul><li>Increasing the speed of insight can help make <strong>faster</strong> and <strong>smarter</strong> <strong>business decisions</strong>.</li><li>Keeping analysts in the analytical flow can <strong>improve efficiency</strong>, leading to <strong>high-quality insights</strong> and analysis.</li><li>Speeding up performance can <strong>reduce dashboard development time</strong>.</li></ul><p>Now it’s clear that the goal is to help users take faster and smarter business decisions while enabling analysts to provide high-quality insights with less time spent on dashboard development.</p><h3>Key Questions Before You Start Developing the Dashboard</h3><p>To achieve that goal, consider the following key questions before you start developing or optimizing the dashboard:</p><ul><li>What is the granularity of the data needed for the dashboard?</li><li>Where is the data source and how is it managed?</li><li>How often does the data need to be updated or refreshed?</li><li>What is the current performance expectation, if any?</li></ul><p>These questions significantly impact how the analyst develops a high-speed dashboard.</p><h3>Understanding Workbook Data Processing Path</h3><p>There are several processes inside the workbook when analysts build the dashboard. Tableau uses visual query language (VizQL) to generate queries to and from data sources. The data processing path has eight steps.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Im_l7hqQz7qAOKhGTVTVDg.png" /><figcaption>Photo from Author</figcaption></figure><h3>How to Run Performance Recording in Tableau?</h3><p>Tableau Desktop includes the performance recording feature to track object load times and identify areas for improvement. Here are the steps to run the performance recording:</p><ol><li>Start from a blank page.</li><li>Click Help, go to Settings &amp; Performance, and choose Start Performance Recording.</li><li>Go to the dashboard that you have developed.</li><li>Repeat step 2, but choose Stop Performance Recording.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/896/1*vmtcKycUSE6JCyS-K4inqQ.png" /><figcaption>Photo from Author</figcaption></figure><blockquote><strong>Pro Tip:</strong> Run your performance recording from a blank page first so that you can get a clean cut of the performance recording. You don’t want anything to render or run a query until you’re ready.</blockquote><h3>How to Read the Performance Result?</h3><p>After you stop the performance recording, you will receive a Tableau Read-Only file to see the performance results. Performance summaries include a <strong>timeline of events</strong>, a <strong>sorted list</strong>, and <strong>any related query details</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0s0SalP76claum4s3zkmyw.png" /><figcaption>Photo from Author</figcaption></figure><p>From the performance recording summary, we can assume that executing query events take a longer time to consume the data in the dashboard. Long-running queries are usually exacerbated by high data volume, complex functions, and view-level calculations like table calculations.</p><p>To help with long-executing queries, you can specify filter fields, remove only relevant values from your filters, replace normal filters with parameters, and set actions.</p><h3>Conclusion Workbook Performance Factors</h3><p>There are four internal factors that affect the dashboard performance, including <strong>Data Design</strong>, <strong>Filtering</strong>, <strong>Calculation</strong>, and <strong>Layout &amp; Visuals</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KK5_3beIb7Qb1UU08LqaWg.png" /><figcaption>Photo from Author</figcaption></figure><p>Data design is the first factor and also represents the biggest opportunity for performance enhancement. Data design can represent the volume of the data and source of filtering. Filtering, Calculation, and Layout &amp; Visuals also affect dashboard performance to varying degrees.</p><p>It would be best to check first how much time the dashboard takes to load your data using performance recording and get the number as the estimated time that stakeholders will face. In the next phase, you can tune the dashboard performance and inform the stakeholders that the dashboard has been optimized from X to Y seconds.</p><p>Feel free to connect with me on <a href="https://www.linkedin.com/in/ariprabowo/">LinkedIn</a>. I would be delighted to share my experiences in the tech industry, data science, marketing, and product.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2a4eb171e737" width="1" height="1" alt=""><hr><p><a href="https://medium.com/cultigo/boosting-tableau-performance-strategies-for-faster-and-more-efficient-dashboards-2a4eb171e737">Boosting Tableau Performance: Strategies for Faster and More Efficient Dashboards</a> was originally published in <a href="https://medium.com/cultigo">Cultigo</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Introduction to Google Script Editor: Automate Your Data Analysis in Google Sheets]]></title>
            <link>https://medium.com/geekculture/introduction-to-google-script-editor-automate-your-data-analysis-in-google-sheets-c654671815c7?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/c654671815c7</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[google-sheets]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Wed, 12 Jul 2023 20:07:13 GMT</pubDate>
            <atom:updated>2023-07-12T20:07:13.036Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2caHoje9qk9mVAsGbjsB5Q.png" /><figcaption>Image from Author</figcaption></figure><h3><strong>Background</strong></h3><p>A couple of months ago (when I wrote this article in June 2023), I was asked to help several colleagues at the company where I work in automating their data processes. Additionally, I have been invited to speak at webinars, campus events, and other online/offline gatherings. The questions I have received are quite similar:</p><ul><li>How can I automate the process of removing duplicate data without manual work?</li><li>If I have more than two sheets and want to compile them into one, how can I automate the process?</li><li>I have fetched data from external sources, how can I retrieve the previous performance (marketing metrics) on a daily basis?</li></ul><p>It is fascinating to see that many individuals are using spreadsheets to automate the process, allowing them to focus on analyzing the data instead of spending a significant amount of time and energy on manual data entry.</p><p>The purpose of this article is to provide you with an understanding of Google Script and its application for data analysis purposes.</p><h3>What is Google Script Editor?</h3><p>The Google Script Editor is a development environment provided by Google for creating custom scripts and automating tasks within various Google Applications, such as Google Sheets, Google Docs, and Google Forms. It allows users to extend the functionality of these applications by writing code based on JavaScript.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y1TJ4imWTSyJxBgLFxivlQ.png" /><figcaption>Image from Author</figcaption></figure><p>In order to access the Google Script Editor in the Google Sheets, you can go to the “Extension” menu, and select “Apps Script”. This will open the Script Editor in a new window or tab.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ik9FKyyCG3tw_21Sdz34iQ.png" /><figcaption>Image from Author</figcaption></figure><h3>Basic Script in Google Script Editor</h3><p>For data analysis purposes, I utilize the Google Script Editor to access and manipulate the specific range of rows and columns in Google Sheets. The following :</p><ul><li>The <strong>`getSheetByName()` </strong>function is used to select the desired sheet for processing.</li><li>The <strong>`getRange()`</strong> function retrieves data from the specified range of rows and columns.</li><li>The <strong>`getLastRow()`</strong> function fetches data up to the last row in the spreadsheet.</li><li>The <strong>`getLastColumn()`</strong> function retrieves data up to the last column in the spreadsheet.</li><li>The <strong>`getValues()`</strong> function retrieves the value from the processed operation.</li><li>The <strong>`setValues()`</strong> function is typically used to set values based on the output of <strong>`getValues()`</strong> or a custom formula</li></ul><blockquote>You can imagine how a script editor works by putting yourself to write the script</blockquote><p>These functions provide crucial capabilities for accessing, manipulating, and analyzing targeted data in Google Sheets. By utilizing the Google Script Editor, data analysts can streamline their workflows and perform advanced data operations efficiently.</p><h4>Moving data from one sheet to another sheet</h4><p>I want you to get your hands dirty by examining the script editor in your Google Sheets. At this part, you will move all data from one sheet to another sheet in the same Google Sheets document.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_u2evA8uh-AUmpPHibRmzw.png" /><figcaption>Image from Author</figcaption></figure><pre>function movingData() {<br>  //1. Get active sheets that you want to process<br>  var sheetOne = SpreadsheetApp.getActiveSpreadsheet().getSheetByName(&quot;Sheet1&quot;);<br>  var sheetTwo = SpreadsheetApp.getActiveSpreadsheet().getSheetByName(&quot;Sheet2&quot;);<br><br>  //2. Get the data from the selected range of cell and get the values<br>  var sheetOneValue = sheetOne.getRange(1,1,sheetOne.getLastRow(), sheetOne.getLastColumn()).getValues();<br><br>  //3. Eliminate the blank rows<br>  var val = sheetOneValue.filter(e=&gt;e[0]); <br>  <br>  //4. From step two and three, put the values in the beginning of the first row and first column in Sheet 2<br>  sheetTwo.getRange(1,1, val.length, val[0].length).setValues(sheetOneValue);<br>}</pre><p>From the provided code, let me explain the step-by-step process of moving data from one sheet to another sheet.</p><ol><li>The variables sheetOneand sheetTwo are used to select the active sheet names, in this case, “Sheet1” and “Sheet2”</li><li>The variable sheetOneValue retrieves the values from the specified range of your cells in Sheet1. The code <strong>`(1, 1, sheetOne.getLastRow(), sheetOne.getLastColumn())`</strong> represents the range of cells from the first row and first column to the last row and last column.</li><li>Step three involves counting the number of filled columns and rows in Sheet1. This step is also used to eliminate any blank rows.</li><li>Finally, the data from Sheet1 will be populated into Sheet2, starting from the first row and columns. The number of columns and rows will be based on the total count of filled data in Sheet1.</li></ol><h4>Formula implementation in Google Script Editor</h4><p>I often receive various requests from my colleagues, including tasks like filtering data based on specific categories or aggregating data. For such requests, the “Query” function in Google Sheets can be a straightforward solution.</p><p>However, what if the scenario involves combining data from multiple sheets into one and filtering out individuals below the age of 30?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OXsBYmufY-xdj5DN60GTdw.png" /><figcaption>Image by Author</figcaption></figure><p>I provide you the view access to <a href="https://docs.google.com/spreadsheets/d/1jh4aS2J_130mTbXxjiyk6TSVH09o3kfgtRf9l_arjus/edit?usp=sharing">Google Sheets</a> to combine several sheets into one sheet and aggregate to filter out the age below 15 years old.</p><pre>function movingData() {<br>  //1. Get active sheets that you want to process<br>  var sheetOne = SpreadsheetApp.getActiveSpreadsheet().getSheetByName(&quot;exampleSheet1&quot;);<br>  var sheetTwo = SpreadsheetApp.getActiveSpreadsheet().getSheetByName(&quot;exampleSheet2&quot;);<br>  var sheetThree = SpreadsheetApp.getActiveSpreadsheet().getSheetByName(&quot;exampleSheet3&quot;);<br>  var sheetCompiled = SpreadsheetApp.getActiveSpreadsheet().getSheetByName(&quot;compiledSheet&quot;);<br><br>  //2. Get the data from the selected range of cell and get the values<br>  var sheetOneValue = sheetOne.getRange(1,1,sheetOne.getLastRow(), sheetOne.getLastColumn()).getValues();<br>  var sheetTwoValue = sheetTwo.getRange(2,1,sheetTwo.getLastRow(), sheetTwo.getLastColumn()).getValues();<br>  var sheetThreeValue = sheetThree.getRange(2,1,sheetThree.getLastRow(), sheetThree.getLastColumn()).getValues();<br><br>  //3. Get rid of blank rows<br>  var valOne = sheetOneValue.filter(e=&gt;e[0]);<br>  var valTwo = sheetTwoValue.filter(e=&gt;e[0]); <br>  var valThree = sheetThreeValue.filter(e=&gt;e[0]); <br><br>  //4. Clear the data in the compiled sheet<br>  sheetCompiled.getRange(1, 1, sheetCompiled.getLastRow(), sheetCompiled.getLastColumn()).clearContent();<br><br>  //5. Compile all data points in each sheet into compiledSheet<br>  sheetCompiled.getRange(1, 1, valOne.length, valOne[0].length).setValues(sheetOneValue);<br>  sheetCompiled.getRange(sheetCompiled.getLastRow()+1, 1, valTwo.length+1, valTwo[0].length).setValues(sheetTwoValue);<br>  sheetCompiled.getRange(sheetCompiled.getLastRow()+1, 1, valThree.length+1, valThree[0].length).setValues(sheetThreeValue)<br><br>  //6. Get the header of the data<br>  var header = sheetCompiled.getRange(1,1,1,sheetCompiled.getLastColumn()).getValues();<br>  <br>  //7. Put the header into another cell and implement aggregation to filter out age above 30<br>  sheetCompiled.getRange(1, 12, 1, 9).setValues(header);<br>  sheetCompiled.getRange(2, 12).setValue(&#39;=QUERY(A2:I67,&quot;SELECT * WHERE A &gt; 30&quot;)&#39;);<br> <br>}</pre><p>The code is quite similar to the previous example to copy and paste from one sheet to another sheet. Additionally, I use a formula to query all data points with a filter of the age below 30 years old.</p><h4>Scheduling your automation</h4><p>In the Google Script Editor, you can schedule your script to automate data analysis tasks using “Trigger”. A Trigger is a special function that allows your script to run automatically based on certain events or time intervals.</p><p>In cases where you need to compile data daily or schedule regular data inputs, I find triggers particularly useful. You can set them to run on a monthly, weekly, or daily basis, or even at intervals as short as minutes. Triggers are flexible and can be adjusted to meet your specific requirements for task automation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kNeZYlRH3KIE9xukwdnJZQ.png" /><figcaption>Image from Author</figcaption></figure><h3>Conclusion</h3><p>When I first started using Google Script Editor, I was amazed at how much time I saved. Tasks that used to take me around 4 hours to compile, process, and aggregate data could now be automated in just 5 seconds using Google Sheets as the primary tool for data analysis.</p><p>The Script Editor is truly remarkable, and I have utilized it in various cases, including product analysis, marketing tasks, operational processes, and growth analysis. It allows me to transform every manual task into automation. Spreadsheets, particularly Google Sheets, have always been my go-to for data handling and analysis. So, why not leverage the power of Google Script Editor?</p><p>I offer a free 15-minute discussion to understand your pain points in using Google Spreadsheet and to help you automate your day-to-day work using Google Sheets. Please follow me on Medium or <a href="https://www.linkedin.com/in/ariprabowo/">LinkedIn</a>, I am happy to share insightful articles that will make your life easier and more productive.</p><p><strong>Thanks for reading!</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c654671815c7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/geekculture/introduction-to-google-script-editor-automate-your-data-analysis-in-google-sheets-c654671815c7">Introduction to Google Script Editor: Automate Your Data Analysis in Google Sheets</a> was originally published in <a href="https://medium.com/geekculture">Geek Culture</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cracking the Code: Unraveling Hypothesis with Powerful Statistical Tests]]></title>
            <link>https://medium.com/geekculture/cracking-the-code-unraveling-hypothesis-with-powerful-statistical-tests-de1fd9f0a082?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/de1fd9f0a082</guid>
            <category><![CDATA[hypothesis-testing]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[statistical-analysis]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Tue, 27 Jun 2023 02:22:11 GMT</pubDate>
            <atom:updated>2025-04-24T21:02:44.944Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*npRcY6LcdJeCVvJTvPuJWg.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@isaacmsmith?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Isaac Smith</a> on <a href="https://unsplash.com/photos/6EnTPvPPL6I?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>A decision is made based on the analysis result of a specific case, problem, or challenge. To validate the findings, conducting a statistical test is essential to ensure the reliability and validity of the conclusions drawn from the data.</p><p>Statistical tests are employed to test hypotheses and determine whether the observed data provide substantial evidence to support or reject a hypothesis.</p><p>The purpose of this article is to discuss commonly used statistical tests in the industry and demonstrate their implementation in Python.</p><h3>Statistical Test and Hypothesis Testing</h3><p>There are several statistical tests that are used in the industry to reveal information or insight, such as t-test, z-test, ANOVA, and ANCOVA to make inferences about the population. Here is a brief explanation of each test:</p><ol><li>T-test: A t-test is used to determine whether there is a significant difference between the means of the two groups. It is commonly used when the sample size is small (typically less than 30) and the population standard deviation is unknown.</li><li>Z-test: A z-test is similar to a t-test but is used when the sample size is large (typically greater than 30) and the population standard deviation is known. It is based on the standard normal distribution and compares the difference between a sample mean and a known population mean.</li><li>ANOVA (Analysis of Variance): ANOVA is a statistical test used to determine whether there are any statistically significant differences between the means of three or more groups. It analyzes the variance between groups and within groups to make this determination.</li><li>ANCOVA (Analysis of Covariance): ANCOVA is an extension of ANOVA that incorporates one or more continuous variables, called covariates, into the analysis. It examines whether there are significant differences in the means of multiple groups while controlling for the influence of the covariates. ANCOVA combines elements of both ANOVA and regression analysis.</li></ol><h3>T-test and Z-test Overview</h3><p>In the t-test and z-test, commonly there are two types of tests: <strong>one-tailed</strong> (or one-sided) and <strong>two-tailed</strong> (or two-sided) tests. The difference lies in the directionality of the hypothesis.</p><p>A one-tailed hypothesis test is used to find that <strong>one group has a difference in a particular direction</strong> from another group. It also specifies a directional difference or effect either a greater-than or less-than relationship. There are two types of one-tailed hypothesis tests, <strong>left-tailed</strong> and <strong>right-tailed</strong>.</p><p>For example, a product manager builds a product/feature on the website to get leads of users into two groups (using form or WhatsApp)</p><h4>Left-tailed hypothesis</h4><p>left-tailed if one group has an effect lower than another group. The left-tailed hypotheses could be:</p><ul><li>Null hypothesis (H0): The effect of WhatsApp brings more leads than the existing method (form)</li><li>Alternative hypothesis (Ha): The effect of WhatsApp on the leads is less than the existing method (form)</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*nDm0dvmwYhZ-E_GG5VXOzg.png" /><figcaption>Photo by <a href="https://statisticsbyjim.com/hypothesis-testing/one-tailed-two-tailed-hypothesis-tests/">Jim Frost </a>on <a href="https://statisticsbyjim.com/hypothesis-testing/one-tailed-two-tailed-hypothesis-tests/">statisticsbyjim.com</a></figcaption></figure><h4>Right-tailed hypothesis</h4><p>Right-tailed if one group has an effect greater than another group. The right-tailed hypothesis could be:</p><ul><li>Null hypothesis (H0): The effect of WhatsApp brings less lead than the existing method (form).</li><li>Alternative hypothesis (Ha): The effect of WhatsApp on the leads is greater than the existing method (form)</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*Er4KuU_WlpiDrHdT7PFfCQ.png" /><figcaption>Photo by <a href="https://statisticsbyjim.com/hypothesis-testing/one-tailed-two-tailed-hypothesis-tests/">Jim Frost </a>on <a href="https://statisticsbyjim.com/hypothesis-testing/one-tailed-two-tailed-hypothesis-tests/">statisticsbyjim.com</a></figcaption></figure><h4>Two-tailed hypothesis</h4><p>Two-tailed hypothesis tests are also known as non-directional, meaning we are testing for a difference or relationship specifying the direction. The two-tailed hypothesis could be:</p><ul><li>Null hypothesis (H0): The mean leads generation of a form and WhatsApp is equal</li><li>Alternative hypothesis (Ha): The mean lead generation using the form is significantly different than using WhatsApp.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*TujnBfWAK8rGl5Koa8lGCw.png" /><figcaption>Photo by <a href="https://statisticsbyjim.com/hypothesis-testing/one-tailed-two-tailed-hypothesis-tests/">Jim Frost </a>on <a href="https://statisticsbyjim.com/hypothesis-testing/one-tailed-two-tailed-hypothesis-tests/">statisticsbyjim.com</a></figcaption></figure><h4>Implementation of t-test and z-test in Python</h4><p>The Python code is quite similar to t-test and z-test, the difference is from the number of samples being used.</p><p><strong>For the t-test:</strong></p><pre>import scipy.stats as stats<br>import numpy as np<br><br># Generate sample data for two groups<br>group1 = np.array([4, 5, 6, 7, 8, 6, 4, 5, 5, 7])<br>group2 = np.array([2, 3, 4, 5, 6, 7, 6, 6, 5, 4])<br><br># One-tailed z-test<br>t_statistic_one_tailed, p_value_one_tailed = stats.ttest_ind(group1, group2, alternative=&#39;greater&#39;)<br>print(&quot;One-tailed t-test p-value:&quot;, p_value_one_tailed)<br><br># Two-tailed z-test<br>t_statistic_two_tailed, p_value_two_tailed = stats.ttest_ind(group1, group2)<br>print(&quot;Two-tailed t-test p-value:&quot;, p_value_two_tailed)</pre><p>Alternative hypothesis and following options (default is ‘two-sided’):</p><ul><li>‘two-sided’: the means of the distribution underlying the samples are unequal.</li><li>‘less’: the mean of the distribution underlying the first sample is less than the mean of the distribution underlying the second sample.</li><li>‘right’: the mean of the distribution underlying the first sample is greater than the mean of the distribution underlying the second sample.</li></ul><p><strong>For the z-test:</strong></p><pre>import scipy.stats as stats<br>import numpy as np<br><br># Generate sample data for two groups (n&gt;30)<br>group1 = np.random.choice(np.arange(1, 31, 1), size=30, replace=False)<br>group2 = np.random.choice(np.arange(1, 31, 1), size=30, replace=False)<br><br># One-tailed z-test<br>z_statistic_one_tailed, p_value_one_tailed = stats.ztest(group1, group2, alternative=&#39;larger&#39;)<br>print(&quot;One-tailed z-test p-value:&quot;, p_value_one_tailed)<br><br># Two-tailed z-test<br>z_statistic_two_tailed, p_value_two_tailed = stats.ztest(group1, group2)<br>print(&quot;Two-tailed z-test p-value:&quot;, p_value_two_tailed)</pre><p>Alternative hypothesis and following options (default is ‘two-sided’):</p><ul><li>‘two-sided’: the means of the distribution underlying the samples are unequal.</li><li>‘smaller’: the mean of the distribution underlying the first sample is less than the mean of the distribution underlying the second sample.</li><li>‘larger’: the mean of the distribution underlying the first sample is greater than the mean of the distribution underlying the second sample.</li></ul><h3>ANOVA Overview</h3><p>ANOVA by definition above is to compare the means of two or more groups to determine if there are any significant differences among them. It analyzes the variation between group means and within-group variation to assess the statistical significance of group differences.</p><p>For example, to determine the significant difference in the revenue from the different numbers of marketing communication sent to the users. The hypothesis could be:</p><ul><li>Null Hypothesis (H0): There is no significant difference between the means of revenue among the number of marketing communication.</li><li>Alternative Hypothesis (Ha): There is a significant difference in revenue means across the various number of marketing communication.</li></ul><h4>Implementation of ANOVA in Python</h4><pre>import statsmodels.api as sm<br>from statsmodels.formula.api import ols<br>import pandas as pd<br><br># Create a sample dataset with multiple groups<br>data = {&#39;NumberCommunication&#39;: [1, 1, 1, 2, 2, 2, 3, 3, 3],<br>        &#39;Value&#39;: [10000, 12000, 14000, 9000, 11000, 13000, 8000, 10000, 12000]}<br>df = pd.DataFrame(data)<br><br># Fit the ANOVA model<br>model = ols(&#39;Value ~ NumberCommunication&#39;, data=df).fit()<br>anova_table = sm.stats.anova_lm(model)<br><br># Print the ANOVA table<br>print(&quot;ANOVA results:&quot;)<br>print(anova_table)</pre><p>In this code, we create a sample dataset with three different of communications and their corresponding values. Then, we use ols a function to specify the model, which includes the dependent variable Valueand the independent variable NumberCommunication . Next, fit the model using fit() to estimate the model parameters.</p><p>The anova_lm function is used to perform the ANOVA analysis and generate an ANOVA table. It calculates the sum of squares, degrees of freedom, F-statistic, and p-value.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/792/1*kxevjRc-ULBeGFMlSBrSbQ.png" /><figcaption>Photo from Author</figcaption></figure><h3>ANCOVA Overview</h3><p>ANCOVA by the definition above is to assess the difference in group means while controlling for the effects of one or more continuous covariates (variable that have an influence on the outcome variable but is not the primary variable of interest).</p><p>For example, researchers want to evaluate the effect of different teaching methods (independent variable) on student’s test scores (dependent variable), while controlling for the student’s pre-test scores (covariate). The hypothesis could be:</p><ul><li>Null Hypothesis (H0): There is no significant difference between the means of test scores among the teaching methods after controlling for pre-test scores.</li><li>Alternative Hypothesis (Ha): There is a significant difference in test scores among the teaching methods after controlling for pre-test scores.</li></ul><h4>Implementation of ANCOVA in Python</h4><pre>import statsmodels.api as sm<br>import pandas as pd<br><br># Test scores for each teaching method<br>scores_method_A = [85, 92, 88, 79, 95]<br>scores_method_B = [75, 82, 80, 81, 78]<br>scores_method_C = [90, 88, 85, 87, 92]<br><br># Covariate variable (prior knowledge)<br>prior_knowledge = [70, 65, 75, 72, 68, 73, 68, 70, 71, 69, 74, 72, 66, 68, 70]<br><br># Combine the data into a dataframe<br>df = pd.DataFrame({&#39;Scores&#39;: scores_method_A + scores_method_B + scores_method_C,<br>                   &#39;Teaching_Method&#39;: [&#39;A&#39;] * 5 + [&#39;B&#39;] * 5 + [&#39;C&#39;] * 5,<br>                   &#39;Prior_Knowledge&#39;: prior_knowledge})<br><br># Step 1: Formulate hypotheses<br># H0: The mean scores of all teaching methods are equal after controlling for the effect of the covariate.<br># H1: At least one of the teaching methods has a different mean score after controlling for the effect of the covariate.<br><br># Step 2: Set significance level (α)<br>alpha = 0.05<br><br># Step 3: Fit the ANCOVA model<br>model = sm.formula.ols(&#39;Scores ~ Teaching_Method + Prior_Knowledge&#39;, data=df).fit()<br>ancova_table = sm.stats.anova_lm(model)<br><br># Step 4: Determine critical value and make a decision<br># Alternatively, you can use the p-value instead of comparing with the critical value.<br>critical_value = stats.f.ppf(1 - alpha, dfn=2, dfd=10)  # dfn = number of groups - 1, dfd = total sample size - number of groups - number of covariates<br>f_value = ancova_table[&#39;F&#39;][0]<br>p_value = ancova_table[&#39;PR(&gt;F)&#39;][0]<br><br>ancova_table = pd.DataFrame(ancova_table)<br>ancova_table</pre><p>From the above code, we create the dataset that contains teaching methods, scores, and pre-test (prior knowledge), then use ols function to specify the ANCOVA model formula. The anova_lm is used to generate an ANCOVA table which contains information about the sources of variation, degrees of freedom, sum of squares, mean squares, F-statistic, and p-value associated with the teaching methods and the covariate.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/748/1*19bTQRzESPz09VaNtNtnJA.png" /><figcaption>Photo from Author</figcaption></figure><h3>Conclusion</h3><p>Statistical tests play a crucial role in data analysis and research. They provide a framework for making evidence-based decisions, drawing valid conclusions, and generalizing findings to populations. Statistical tests enable hypothesis testing, allowing analysts or researchers to evaluate the strength of evidence for or against a specific hypothesis.</p><p>Overall, statistical tests contribute to scientific rigor, validity, and reliable research findings across diverse fields of study. Have you applied statistical tests in your analysis?</p><p>Feel free to connect with me on <a href="https://www.linkedin.com/in/ariprabowo/">LinkedIn</a>. I would be delighted to share my experiences in the tech industry, data science, marketing, and product.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=de1fd9f0a082" width="1" height="1" alt=""><hr><p><a href="https://medium.com/geekculture/cracking-the-code-unraveling-hypothesis-with-powerful-statistical-tests-de1fd9f0a082">Cracking the Code: Unraveling Hypothesis with Powerful Statistical Tests</a> was originally published in <a href="https://medium.com/geekculture">Geek Culture</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Unleash the CLTV Secret: Skyrocket Your Marketing ROI with High-Value Customers]]></title>
            <link>https://medium.com/geekculture/unleash-the-cltv-secret-skyrocket-your-marketing-roi-with-high-value-customers-cf76cb5a28bb?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/cf76cb5a28bb</guid>
            <category><![CDATA[customer-lifetime-value]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Wed, 14 Jun 2023 01:20:10 GMT</pubDate>
            <atom:updated>2023-06-14T01:48:40.326Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QlEvaT6M8vmgW3eIeirJFQ.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@blakewisz?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Blake Wisz</a> on <a href="https://unsplash.com/photos/tE6th1h6Bfk?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>Customers who are acquired through marketing should be assessed to determine whether they can bring value within a specific timeframe. You may have the same questions as me:</p><ul><li>Do I acquire high-value customers?</li><li>Do I know when I will receive a return on my investment after acquiring new customers?</li><li>How much did I spend on acquiring new customers based on last year’s customer acquisition data?</li></ul><p>The purpose of this article is to assist businesses to identify high-value customers, effectively allocating marketing resources, and developing strategies to maximize customer retention.</p><h3>CLTV and Its Advantages</h3><p>CLTV (Customer Lifetime Value) is an essential task for enterprises as it helps identify the value of customers throughout the business relationship. There are several advantages to conducting CLTV analysis:</p><ul><li><strong>Customer acquisition and retention</strong>, CLTV analysis assists businesses in understanding the profitability of acquiring and retaining customers. By focusing on acquiring similar customers, businesses can maximize customer value and loyalty.</li><li><strong>Marketing effectiveness</strong>, CLTV analysis evaluates the effectiveness of channels and campaigns. This enables businesses to optimize their marketing strategies, leading to better Return on Investment (ROI)</li><li><strong>Customer segmentation</strong>, CLTV analysis aids businesses in segmenting customers based on similar CLTV ranges. This allows for tailored marketing, product offerings, and customer service that cater to the unique needs and preferences of each segment</li></ul><p>Here is the common calculation of CLTV:</p><blockquote>CLTV = (Average Purchase Value) x (Purchase Frequency) x (Customer Lifespan)</blockquote><p>Regarding the formula, CLTV can be further refined with additional factors such as customer acquisition cost, and customer churn rate. These factors provide a more accurate estimate of the long-term profitability associated with customers.</p><blockquote>Disclaimer: The example and problem-solving for the rest of this article will use a cohort analysis approached</blockquote><h4>Customer Lifespan</h4><p>Customer lifespan refers to the duration during which a customer actively engages with a business, making purchases or utilizing its products or services. This duration is typically measured in months or years, depending on the specific business and industry. To illustrate the concept of customer lifespan for CLTV analysis, let’s consider an example based on cohort analysis:</p><p>Cohort A: Customers acquire in January 2020</p><ul><li>Customer 1: First purchase on January 01, 2020; Last purchase on December 30, 2020 (12 months)</li><li>Customer 2: First purchase on January 15, 2020; Last purchase on October 15, 2020 (10 months)</li><li>Customer 3: First purchase on January 21, 2020; Last purchase on November 30, 2020 (11 months)</li></ul><p>To calculate the customer lifespan, you need to measure the duration between the first purchase and the last purchase for the customers in each cohort. Here is the lifespan of cohort A:</p><ul><li><strong>Cohort A</strong>, the lifespan ranges from 10–12 months, with an <strong>average of around 11 months</strong></li></ul><h4>Purchase Frequency &amp; Average Purchase Value</h4><p><strong>Purchase frequency</strong>, also known as a purchase rate or buying rate, measures the number of times a customer place orders within a specific period. For example, if customer A makes 100 purchases in a year, the purchase frequency would be calculated as 100 purchases divided by 1 year, resulting in a purchase frequency of 100 purchases per year. Let’s delve into the calculation of purchase frequency:</p><blockquote>Purchase frequency = Total Number of transactions / customer lifespan</blockquote><p>For instance, let’s consider three customers with transaction counts of 10, 18, and 15 transactions, the customers have utilized the service or product for around 1 year and a half, resulting in an average purchase frequency of approximately 28 transactions per 1 year and a half.</p><p><strong>Average purchase value</strong>, also known as average order value (AOV), represents the average monetary value (selling price) of individual purchases within a specific period. Typically, businesses include the total spending of the company for each transaction, such as the buying price of the products. The calculation for determining the average purchase value is as follows:</p><blockquote>Average Purchase Value =</blockquote><blockquote>(Total selling price-Total buying price) / Total number of transactions</blockquote><p>For instance, let’s consider customer A who has conducted 100 transactions, resulting in a total revenue of approximately $3,000. Additionally, the business incurs a cost of $500 to send orders to customers. Therefore, the calculation for the average purchase value would be ($3,000 — $500) / 100, which equates to an average purchase value of around $25.</p><h4>Practical Example of CLTV</h4><p>Let’s calculate the CLTV for the dream company. The company intends to analyze the acquisition that took place in January 2020. Their objective is to determine the amount of money that should be spent in February to achieve profitability. Here are the step-by-step instructions for calculating the CLTV based on the January 2020 acquisition:</p><p><strong>Customer acquisition in January 2020, </strong>with the <strong>Cost of Acquisition </strong>(<strong>CAC) of around $200</strong></p><ul><li><strong>Customer 1</strong><br>- Lifespan: 12 months<br>- Purchases: 20<br>- Total revenue: $440<br>- Company cost: $53</li><li><strong>Customer 2</strong><br>- Lifespan: 13 months<br>- Purchases: 14<br>- Total revenue: $450<br>- Company cost: $62</li><li><strong>Customer 3</strong><br>- Lifespan: 11 months<br>- Purchases: 10<br>- Total revenue: $320<br>- Company cost: $20</li></ul><p>Let’s take all metrics and calculate each variable to get CTLV:</p><ul><li>Cohort lifespan ranges from 11 to 13 months, with an <strong>average lifespan of approximately 1 year</strong></li><li>Average purchase value <br>= (($440–$53)+($450–$62)+($320-$20)) / (20+14+10 purchases)<br>= <strong>$24.4 per purchase</strong></li><li>Purchase frequency<br>= (20+14+10) / 1 year<br>= <strong>44</strong> <strong>purchases per year</strong></li><li>CLTV for customers that acquire in January 2020<br>= $24.4 per purchase * 44 purchase / year * 1 year<br>= <strong>$1,073<br>= $357.8 per customer</strong></li></ul><p>In the practical example of CLTV, you have evaluated the CLTV based on the cohort acquisition in January, with a CAC of approximately $200. The result indicated that the CLTV amounts to $357.8 from the customer in the following year. In order to get profitability for the next year, you might not spend more than the CLTV, your acquisition cost should be lower than CLTV.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kCRGUruY5y8_XmleVhzfxw.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/fr/@alexandermils?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Alexander Mils</a> on <a href="https://unsplash.com/photos/lCPhGxs7pww?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>Conclusion</h3><p>CLTV is a crucial metric that estimates the total value a customer brings to a business over their entire relationship. It is an important concept for businesses as it provides insight into the long-term profitability of customers and helps the business to create a robust strategic decision related to customer acquisition, retention, and marketing efforts.</p><p>The above explanations with practical examples might be your fundamental study to utilize with various cases such as CLTV for the specific channel of marketing like Facebook ads, Google ads, and programmatic channels.</p><p>Now it is your turn to apply CLTV in your business and organization, talk to your marketing manager and work with a data analyst to assess your business value.</p><p>Feel free to connect with me on <a href="https://www.linkedin.com/in/ariprabowo/">LinkedIn</a>. I would be delighted to share my experiences in the tech industry, data science, marketing, and product.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cf76cb5a28bb" width="1" height="1" alt=""><hr><p><a href="https://medium.com/geekculture/unleash-the-cltv-secret-skyrocket-your-marketing-roi-with-high-value-customers-cf76cb5a28bb">Unleash the CLTV Secret: Skyrocket Your Marketing ROI with High-Value Customers</a> was originally published in <a href="https://medium.com/geekculture">Geek Culture</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Optimizing Your Workload: The Scrum Methodology for Your Team’s Productivity]]></title>
            <link>https://medium.com/geekculture/optimizing-your-workload-the-scrum-methodology-for-your-teams-productivity-2b998771498a?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/2b998771498a</guid>
            <category><![CDATA[teamwork]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[scrum]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Tue, 30 May 2023 01:13:00 GMT</pubDate>
            <atom:updated>2023-05-30T01:13:00.806Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QLH43273adgcaMN_m1l6iA.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@lazizli?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Lala Azizli</a> on <a href="https://unsplash.com/s/photos/scrum-methodology?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>Maintaining your team’s productivity creates a healthy work environment, enabling efficient work execution, skill enhancement, and valuable contributions to your company’s success. Consider the following questions:</p><ul><li>Has your leader encouraged you to take on tasks that promote personal growth?</li><li>Do you find it challenging to prioritize tasks and ensure transparency with stakeholders?</li><li>Are there measurable metrics associated with your assigned tasks?</li><li>Do you regularly conduct retrospectives with your team to discuss projects and identify areas for improvement?</li></ul><p>The purpose of this article is to guide you and your team toward embracing the Scrum methodology, using my personal experience as an example.</p><h3>The Scrum Methodology Framework</h3><p>According to <a href="https://www.scrum.org/">Scrum.org</a>, Scrum is primarily used by tech teams consisting of a Scrum Master, Product Manager, and Developers. However, it is adaptable for various teams, providing a clear structure for collaboration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1RHckilkQ-cV8HQVXuO1Wg.png" /><figcaption>Photo by Ken Schwaber on <a href="https://www.scrum.org/about">Scrum.org</a></figcaption></figure><p>Several tools can help implement Scrum in your team, such as Jira Scrum, Trello, and other platforms. In my example, I will share my experience using <strong>Scrum Kanban</strong> and additional tools like <strong>Miro</strong></p><h3>Roadmap &amp; Product Backlog</h3><p>During this stage, your leader will share the projects your team will undertake. These projects are often referred to as “Epics” or main projects. For instance:</p><ul><li>A/B testing for the action button on the checkout page</li><li>User behavior tracker in the article page to assess product feasibility</li><li>Machine learning model to classify search keywords into relevant product categories</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ErudsZj4ndZVkJFvK2tY4g.png" /><figcaption>Screenshot from the Scrum Jira Software</figcaption></figure><p>To achieve these projects, team members must be assigned to specific Epics. Each Epic consists of several tasks, which should be transparently organized in a backlog.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*x48dukYQHektWtkIHDfq0g.png" /><figcaption>Screenshot from the Scrum Jira Software</figcaption></figure><h3>Sprint Planning</h3><p>In this stage, you and your leader discuss the desired outcomes of the sprint. For example, increasing the click-through rate by 50% and transactions by 20% for product recommendations from search results.</p><p>During sprint planning, your leader will ask questions related to the tasks:</p><ul><li>When should each task start and finish?</li><li>What is the workload estimation in terms of story points?</li><li>How should the tasks be prioritized?</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ujIxS7HdEQgiSx4bVQvPAg.png" /><figcaption>Screenshot from the Scrum Jira Software</figcaption></figure><h4>Story Point</h4><p>Story points are essential metrics that indicate the effort required to complete a task. They typically follow a Fibonacci sequence, with a maximum agreed upon by the team (e.g., 5). Here’s an example of story point estimation used in my team:</p><ul><li>1 story point: Can be completed in 1–2 hours</li><li>2 story points: Can be completed in 3–4 hours</li><li>3 story points: Can be completed in 4–6 hours</li><li>5 story points: Can be completed in up to 8 hours</li></ul><blockquote>Maintaining a balance between task execution and skill development should be a priority for your team’s well-being</blockquote><p>In my team’s experience, each person consumes around 40 story points per sprint, with a tolerance of +5 story points. Remember to set realistic thresholds based on your team’s capacity.</p><h3>Sprint Backlog</h3><p>Sprint backlog is used to say as a sprint prioritization in my team. The reason is to ensure your leader manages the backlog, adjusting task priority if needed, and allowing a buffer of 5–10 story points for urgent tasks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J_Mmn5TEYGHMZJhIMvVyqg.png" /><figcaption>Screenshot from the Scrum Jira Software</figcaption></figure><p>Once the leader has managed properly and is ready to do the project, then the leader starts the sprint and later the leader will follow up on the progress in the <strong>Daily Scrum</strong> stage.</p><h3>Daily Scrum</h3><p>Daily Scrum meetings don’t necessarily have to be daily stand-ups. Adapt them to your organization and team. In my experience, we have two meetings per week and use communication platforms for progress updates on the other three days.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LaAhyWCI07kXxjezNQfCsQ.png" /><figcaption>Screenshot from Google Calendar</figcaption></figure><p>During the 15-minute daily stand-up, share progress updates and address any blockers. Examples of discussions may include:</p><ul><li>The machine learning model has been saved as a pickle file and is ready for integration into the API process.</li><li>The data automation in BigQuery is malfunctioning, hindering data retrieval for sales trend analysis. I will contact the data engineer to resolve the issue.</li></ul><h3>Sprint Review and Sprint Retrospective</h3><p>The final stages of the Scrum methodology are the sprint review and retrospective. In my team, we allocate over 1.5 hours for a meeting at the end of each sprint (usually on Fridays). The meeting includes:</p><ul><li>Spending 45–60 minutes summarizing completed projects or tasks from the sprint.</li><li>Allocating 15 minutes to review tasks, calculate priorities, and story points.</li><li>Dedicate 15–30 minutes for a retrospective, where each person shares what worked well, areas for improvement, and suggestions for enhancement.</li></ul><blockquote>The more team members you have, the more time will be required for meetings. Remember to prioritize the productivity of your team during these meetings.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/785/1*y1SyPkoD7SlyUltrrnakZw.png" /><figcaption>Retrospective Boards in Miro</figcaption></figure><h3>Conclusion</h3><p>Implementing the Scrum methodology can significantly impact your team’s transparency, workload management, and productivity. By planning projects with clear objectives, assigning tasks with appropriate workload allocation and prioritization, regularly communicating progress and blockers, and conducting sprint reviews and retrospectives, you can unlock your team’s true potential, and benefit both individuals and the company.</p><p>Now, it is your turn. Are you ready to embrace the Scrum methodology and create a productive work environment? Engage in discussions with your colleagues and leaders, brainstorm ideas, and tailor the methodology to suit your team’s working style.</p><p>Feel free to connect with me on <a href="https://www.linkedin.com/in/ariprabowo/">LinkedIn</a>. I would be delighted to share my experiences in the tech industry, data science, marketing, and product.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2b998771498a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/geekculture/optimizing-your-workload-the-scrum-methodology-for-your-teams-productivity-2b998771498a">Optimizing Your Workload: The Scrum Methodology for Your Team’s Productivity</a> was originally published in <a href="https://medium.com/geekculture">Geek Culture</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My Story to Get 1st Place in Dataviz Competition]]></title>
            <link>https://medium.com/cultigo/my-story-to-get-1st-place-in-dataviz-competition-a8fd0b463530?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/a8fd0b463530</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[data-analytics]]></category>
            <category><![CDATA[tableau]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Tue, 02 May 2023 00:51:45 GMT</pubDate>
            <atom:updated>2023-05-02T00:51:45.273Z</atom:updated>
            <content:encoded><![CDATA[<p>Doing research is the most activity that you have done to stand out from other participants.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qDpqY2CybPxKGTkqw_zIng.jpeg" /><figcaption>Awardee Session in Jabar Dataviz Festival 2022</figcaption></figure><p>The objective of this article is to remind me as well as to share my thoughts to you to always put the problem, customers (in business) as the priority that you want to solve. Hopefully, it helps you whoever:</p><ul><li>Want to pursue a career in data (especially data visualization)</li><li>Get a chance to be a winner in the data visualization competition</li><li>Enhance your skills in the best practice of data visualization in your work</li></ul><blockquote><strong>Disclaimer:</strong> it is not a straightforward process to get a winning in the competition, the iteration is always the key to maximizing the output.</blockquote><h3>Doing research to understand the status quo of the topic</h3><p>Around 50% of the overall activities in this competition, I dedicate my time to do research with my team. I started asking several questions such as:</p><ul><li>What West Java tends to be like in the future?</li><li>Why the issue is raised by the government?</li><li>What kind of plans that the government in West Java have?</li><li>Why the society needs those plans to make life easier in West Java?</li><li>How the plans can be amplified with the initiative that we (as a society) propose to help the government?</li></ul><p>Here are what I searched to support the status quo of the topic:</p><ul><li>Ridwan Kamil Kebut 11 Program Prioritas di Jabar 2022, CNN Indonesia</li><li>Laporan Perekonomian Provinsi Jawa Barat Mei 2022, Bank Indonesia</li><li>Petani Rusak Tanaman Karena Harga Sayur Anjlok, Dinas Pertanian Kabupaten Bandung Beri Tanggapan, bandung.Kompas.com</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/650/1*oNoJ8f1hOOXijACpzBJ0oQ.jpeg" /><figcaption>Governor of West Java Ridwan Kamil signed KUA-PPAS 2022 at the DPRD office of West Java — Archives of the West Java Provincial Government</figcaption></figure><p>Of course, the research must bring the output to not only criticize the government but also give our hand to make the plan succeed.</p><h3>Data storytelling that follows the chart rule</h3><p>Around 25% of my time exploring what kind of data as well as the chart I must build to catch the eye of the audience in the first place. In the past two years, I focus and invest a lot in data visualization, particularly data storytelling. It makes the audiences understand the insight and summarize the information that you are telling before explaining it to them.</p><blockquote>Simply showing your data without telling the story, it brings headache to the audience</blockquote><p>The first step in data storytelling is the selection of your chart. There is plenty of charts that you can use to visualize the data and turn it into information or insight, but you must be aware of the purpose of each chart. There are four categories in building your chart from your data:</p><ol><li>Comparison, to compare the magnitude of value to each other</li><li>Distribution, to see how quantitative values are distributed and how often each value occurs in the dataset</li><li>Composition, to see how a part compares to the whole data</li><li>Relationship, to see the relationships between two or more variables</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7LUB0i9r-lc6Cq-UpXPbgQ.jpeg" /><figcaption>Starter Chart Visualization, Photo by <a href="https://www.techprevue.com/decision-tree-perfect-visualisation-data/">Techprevue</a></figcaption></figure><h3>Your solution also includes action items</h3><p>The problem has been stated above when you do your research. Therefore, the end of your story should contain any action items. The action items could be the campaign to remind every stakeholder to achieve the goal or to encourage others to work together to achieve the goal.</p><blockquote>Touch your audiences with the problem for all of us, and the audience (especially judges) will love you if you involve them to solve the problem</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sca0kYZBUoY4HcRnnRjvoQ.jpeg" /><figcaption>Giving a presentation to cultivate plants using hydroponic system</figcaption></figure><p>The action item might be difficult to be found and prove that it will solve the problem. However, that is the challenge that you have to encounter to be a champion. Fortunately, I have several things that might help you to find the action item:</p><ul><li>Involving yourself in volunteer activities like helping farmers with updated skills in rural areas.</li><li>During your volunteer, you will get a bunch of solutions that came from your fellow volunteer.</li><li>Summarize your activity with the successful solution along with the problem that you face.</li></ul><blockquote>Simplify your action item, always think that it will solve the problem, no matter if it brings small or big changes</blockquote><p>There are so many problems that you will try to solve, no rush to solve all of them, just pick one or two problems and solve them with your best. It must be valuable for those who need the solutions.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a8fd0b463530" width="1" height="1" alt=""><hr><p><a href="https://medium.com/cultigo/my-story-to-get-1st-place-in-dataviz-competition-a8fd0b463530">My Story to Get 1st Place in Dataviz Competition</a> was originally published in <a href="https://medium.com/cultigo">Cultigo</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Applied Group, Set and Parameter for Report and Dashboard in Tableau]]></title>
            <link>https://medium.com/geekculture/applied-group-set-and-parameter-for-report-and-dashboard-in-tableau-6e401873031f?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/6e401873031f</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[data-analytics]]></category>
            <category><![CDATA[storytelling-for-business]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Mon, 08 Nov 2021 06:32:14 GMT</pubDate>
            <atom:updated>2021-11-08T06:32:14.621Z</atom:updated>
            <content:encoded><![CDATA[<p>An ultimate handy dynamic filter for your analysis</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*n6g4YpKEM5yq1Ipz4CpnTg.jpeg" /><figcaption>Image by <a href="https://unsplash.com/@claybanks">Clay Banks</a> from <a href="https://unsplash.com/photos/_Jb1TF3kvsA">Unsplash</a></figcaption></figure><blockquote>For those of you who are just starting to analyze data using Tableau, this article will help you to understand one of the basic features in Tableau. For those of you who are familiar with Tableau, let’s refresh our lessons.</blockquote><h4>Contents</h4><p>There are four things that you will get in this article:</p><ol><li><strong>Group</strong>, to combine related members in a field</li><li><strong>Set</strong>, to compare the selected data and others</li><li><strong>Parameter</strong>, to replace the constant value with your determined threshold</li><li><strong>Exercise</strong> to utilize Group, Set &amp; Parameter to solve the business problem</li></ol><h4>First thing first</h4><p>Before understanding about group, set, and parameter. I give credit to <a href="https://www.linkedin.com/in/keremenko/"><em>Kirill Eremenko</em></a> and the team for giving the challenge to find any potential startups to be invested in (dummy data) which can be applied to any similar cases outside there. You can download the dataset in <a href="https://github.com/densaiko/data_science_learning/blob/main/dataset/Startups%20Data.xlsx"><strong>my repository</strong></a>, and here is the challenge.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sYEzPX6Vid1grQjodQonLQ.png" /><figcaption>Image is created by Author</figcaption></figure><h4>Group</h4><p>In Tableau, group is used to combine related members in a field. For instance, you have small data points from 1999 to 2008, and instead of putting them in the visualization every year, you decide to put them in one member called “1999 to 2008”.</p><p>I will show you using <a href="https://github.com/densaiko/data_science_learning/blob/main/dataset/Startups%20Data.xlsx"><strong>startup dataset</strong></a>, here are the steps:</p><ol><li>In the data pane, right-click on <strong>Year Founded </strong>and select <strong>create &gt; group</strong></li><li>Block the year 1999 to 2008 and click <strong>Group</strong></li><li>Naming the group and clicking <strong>Ok</strong></li><li>Put <strong>field Group</strong> into <strong>columns</strong> and <strong>metrics</strong> into <strong>Rows</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7HU4-P_E2cBAN05n3ZOLmQ.png" /><figcaption>Image is created by Author</figcaption></figure><p>The result that you will get is down below</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AOe_dj1KAtgH_eke5S5wig.png" /><figcaption>Image is created by Author</figcaption></figure><h4>Set</h4><p>In Tableau, set is used to compare the selected data and turn it into subset data. For instance, the board of directors needs to see find the top 10 growth startups in 2015. Here are the steps to see the top 10 growth leaders:</p><ol><li>In the data pane, right-click the field <strong>name</strong> and select <strong>create</strong> &gt; <strong>set</strong></li><li>Fill in the name “Growth Startup” and click <strong>use all</strong> data</li><li>Choose <strong>Top</strong> &gt; click <strong>by field</strong> &gt; fill <strong>10</strong> &gt; Choose <strong>2015 Growth %</strong></li><li>Put field <strong>Growth Startup </strong>into the <strong>filter box</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iJBo7fMsHeOavb--i8g8yQ.png" /><figcaption>Image is created by Author</figcaption></figure><p>The result that you will get is down below</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*U6z-CUmCXwTMMfMzO-Vg5g.png" /><figcaption>Image is created by Author</figcaption></figure><h4>Parameter</h4><p>In Tableau, parameter is used to replace a constant value in a calculation with your determined threshold. For instance, the board of directors wants to see a top N growth startup and needs a live filter to do that. Here are the steps to do that:</p><ol><li>In the data pane, right-click the field <strong>name</strong> and select <strong>create</strong> &gt; <strong>parameter</strong></li><li>Fill the name “2015 Growth % Startup Parameter” and fill in the properties like in the picture</li><li>Go to field “Growth Startup” and <strong>edit set</strong></li><li>Change the top 10 into “2015 Growth % Startup Parameter” and ok</li><li>On the field parameter &gt; right click and <strong>show parameter</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C5-efs9k6HTuRxlsccnwvA.png" /><figcaption>Image is created by Author</figcaption></figure><p>The result that you will get is down below</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bOwTqG0Dkmk59l6W7wTKXQ.png" /><figcaption>Image is created by Author</figcaption></figure><p>Congratulations, now you understand several basic concepts that Tableau has. Now, it’s time to utilize them to solve the problem. Here is the request that the board of directors wants from you:</p><ul><li>An intended filter of cutoff revenue</li><li>An intended filter of cutoff expense</li><li>Show them in quadrant visualization</li></ul><p>To do them, let’s make it simple by seeing the flow below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mHZHLFeIYTt7Fb6-i37A7Q.png" /><figcaption>Image is created by Author</figcaption></figure><p>The result of the startup quadrant should be like this</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BQDiMXvJ6EDPLJzdR7ZnSQ.png" /><figcaption>Image is created by Author</figcaption></figure><h4>Final Touch</h4><p>Now, you are ready to put all of them into one in the dashboard. You can go to the new dashboard and put the <strong>Growth Startup </strong>sheet and <strong>Startup Quadrant</strong> sheet in your dashboard. The result should be like this</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sfJFsFbQa2CcU34rOlPPQQ.png" /><figcaption>Image is created by Author</figcaption></figure><h4>Conclusion</h4><p>You may face some difficulties parts in this tutorial, let’s make it your exploration because I believe you will get more whether to add some features or other analysis. You can find the Tableau viz <a href="https://public.tableau.com/app/profile/ari.sulistiyo.prabowo/viz/StartupAnalysis_16361761602700/StartupQuadrant">here</a>. I hope you enjoy this and please follow me on my <a href="https://www.linkedin.com/in/ariprabowo/"><strong>LinkedIn</strong></a> and <a href="https://github.com/densaiko"><strong>Github</strong></a>. I am happy to get any positive feedback for future articles. Happy learning!!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6e401873031f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/geekculture/applied-group-set-and-parameter-for-report-and-dashboard-in-tableau-6e401873031f">Applied Group, Set and Parameter for Report and Dashboard in Tableau</a> was originally published in <a href="https://medium.com/geekculture">Geek Culture</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[You Should Have Cold Money to Run Your Side Hustle]]></title>
            <link>https://medium.com/fortune-for-future/you-should-have-cold-money-to-run-your-side-hustle-eb548e8094f3?source=rss-125434829b03------2</link>
            <guid isPermaLink="false">https://medium.com/p/eb548e8094f3</guid>
            <category><![CDATA[side-hustle]]></category>
            <category><![CDATA[money]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Ari Sulistiyo Prabowo]]></dc:creator>
            <pubDate>Wed, 26 May 2021 12:00:41 GMT</pubDate>
            <atom:updated>2021-05-26T12:00:41.721Z</atom:updated>
            <content:encoded><![CDATA[<h4>You are preparing for yourself in the future with cold money</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BZZY-k2IKP1IScaNjZGy1w.jpeg" /><figcaption>Image by <a href="https://unsplash.com/@giorgiotrovato">Giorgio Trovato</a> from <a href="https://unsplash.com/photos/BRl69uNXr7g">Unsplash</a></figcaption></figure><p>In the context of money, I have been separated which money (cold money) that I use to multiple it deliberately even if I will lose during the journey.</p><p>I knew the concept of hot and cold money since I learned about investment. What I always remember until right now is</p><blockquote>In the stock market, you can use your cold money to have a capital gain. Even if you lose, but you are okay with that because you already fulfill your daily needs with hot money</blockquote><p>However, cold money is not only used in the investment but it can be used widely. For instance, I use it to start my own business since I was in college which I earn $2000 of extra income and earn 40–100% capital gain in the stock market.</p><h3>First and foremost</h3><p>Everyone must realize how much money you get should be separated into two different parts at least. The reason is that we can <strong>secure part of our money</strong> and/or <strong>let them work to get more money</strong>.</p><p>Yes, there are two key points above it. What does that mean?</p><p><strong>“Secure”</strong> means I want my money to stand up against inflation. On the other hand, “<strong>let them work” </strong>means I want my money not only to stand up against inflation but stand up to get more beyond inflation.</p><p>So, I decided to put my main income into two parts.</p><ol><li><strong>Hot money,</strong> I separate this money in order to fulfill my daily life such as housing, food, transportation, and others related to my needs. The portion for hot money depends on your monthly expense. For my side, the portion is 40% for the hot money.</li><li><strong>Cold money</strong>, this money is where I use it to grow my skills if I want to go to the next level job or promotion, I also use it to start my business, and some of the cold money is used to invest it such as stock market. If the portion of hot money is 40% then my portion in cold money is 60%.</li></ol><p>Well, after reading until this part, I suggest you take a note and calculate</p><ul><li>How much money do you earn in a month?</li><li>What daily needs that you have to buy/pay in a month (please be wise to it and <strong>only needs</strong> not wants)</li><li>Then you calculate, what portion do you have between hot money and cold money</li></ul><blockquote>Understanding your expense right now is a kind of way to understand and care for yourself in the future</blockquote><h3>Cold Money Management</h3><p>Congratulation! Now you have secured your money without losing it too much by separating them into two parts.</p><p>In this section, I will talk further about cold money management.</p><blockquote>It is because you have secured your daily life with your hot money, it does not mean you can do what you want to do with your cold money without maintaining it.</blockquote><p>Let’s come back to the example of mine with a 60% portion of my money as cold money.</p><p>From 100% of my cold money, I separate it again into several parts (it depends on your journey as a side hustle). I separate it into four parts:</p><ol><li>Startup (Hidroagro) — 40%</li><li>Investment (stock market mostly) — 20%</li><li>Skills — 20%</li><li>Cash — 20%</li></ol><h4>Startup</h4><p>I like spending my energy and time to think how to solve the problem in society. There is satisfaction when you solve their problem and they are happy with your product, and you also happy that you are worth it in society. Therefore, I start my own business called Hidroagro. Before Covid-19, I earn $2000 profit from my business. Now, I continue again after being vacuum because of Covid-19.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MedRqD3MBh-pnPoH-xccLQ.jpeg" /><figcaption><a href="https://www.instagram.com/hidroagro_/?hl=en">Hidroagro</a>, My Startup by Author</figcaption></figure><h4>Investment</h4><p>The cold money jargon comes after learning about investment when I was in college. At that time, I was in the second year of my study in nuclear engineering. But I know after reading a book that we can’t afford to do a 9–5 job until old. Therefore, I start to invest my money around 20% in the stock market and get good capital gain.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jieDsZSX-0mdNvBGW_ENDg.png" /><figcaption>Image is created by Author</figcaption></figure><h4>Skills</h4><p>I believe that every three months you have to add more skills to enrich your ability to solve any problem in your job, business, and life. Why three months? Because I need one month to learn and two months to repeat. The repetition is the strongest one to remember how you utilize the skills. Anyway, this is my way to enrich my ability. You can do it at your own pace and have fun doing it, that is the most important to feel the journey.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*dbY69drWdCxyZuhm_U8jcg.png" /><figcaption>Image from Author</figcaption></figure><h4>Cash</h4><p>Sometimes I overthink about three points above that how if they can’t give me a good return or the skills that I learn is not suitable with the case today. But, I think it is a kind of human nature to defend when there is a bad situation appear. Therefore, I still keep my cash from my cold money.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BZZY-k2IKP1IScaNjZGy1w.jpeg" /><figcaption>Image by <a href="https://unsplash.com/@giorgiotrovato">Giorgio Trovato</a> from <a href="https://unsplash.com/photos/BRl69uNXr7g">Unsplash</a></figcaption></figure><h3>The Advice</h3><p>Based on the title, I will give you my advice to manage your money on how to prepare for your side hustle in the future.</p><p>I know that it is not easy to start a business, spare your money to invest, or maybe you have no money to get new skills. Therefore, you have to understand where are you right now and what you have today. You have to aware of your foundation.</p><ol><li>Calculate your monthly expense (wise to calculate it and only for needs not wants) and after that separate your money into hot money and cold money</li><li>When you start a business, investing, or skills. <strong>You just remember that you are ready to lose it</strong> because starting a business will consume a lot of money and time in the first year, and lose some of the money with the wrong decision to choose stocks and others. However, you still have 20% of the cash and 20% of the skills that you can use to get more money.</li><li>Your business runs well and your portfolio shows the green percentage (capital gain), you have to remember that you need to <strong>reinvest</strong> from what you get from them in your business and investment. Because you can’t afford to work until you are old.</li></ol><p>These three pieces of advice are enough for us. You simply remember, implement and discipline. I believe it will help you to achieve your financial freedom before you get retired.</p><p>I hope this article is helpful and you can prepare yourself before starting your side hustle. If you think this article is a gold article, please follow me, clap, and comment down below. So, I will write another article related to this article.</p><p>This article is based on my experience on how to allocate my money. It should not be considered Financial or Legal Advice. My experience will not be accurate with the general conditions and your capability. Therefore, you may need to calculate it with your condition before making any significant financial decisions.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eb548e8094f3" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fortune-for-future/you-should-have-cold-money-to-run-your-side-hustle-eb548e8094f3">You Should Have Cold Money to Run Your Side Hustle</a> was originally published in <a href="https://medium.com/fortune-for-future">Fortune For Future</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>