<?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 Adeleye, Theophilus Ademola Oluwaferanmi on Medium]]></title>
        <description><![CDATA[Stories by Adeleye, Theophilus Ademola Oluwaferanmi on Medium]]></description>
        <link>https://medium.com/@theophilusadeleye?source=rss-fadf440416cd------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*dmbNkD5D-u45r44go_cf0g.png</url>
            <title>Stories by Adeleye, Theophilus Ademola Oluwaferanmi on Medium</title>
            <link>https://medium.com/@theophilusadeleye?source=rss-fadf440416cd------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 14:49:51 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@theophilusadeleye/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[Never Trust a Number You Can’t Trace: My Week Painting Data with Pandas and Seaborn]]></title>
            <link>https://medium.com/@theophilusadeleye/never-trust-a-number-you-cant-trace-my-week-painting-data-with-pandas-and-seaborn-628f15a03e8c?source=rss-fadf440416cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/628f15a03e8c</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[pandas]]></category>
            <category><![CDATA[internships]]></category>
            <category><![CDATA[data-scientist]]></category>
            <dc:creator><![CDATA[Adeleye, Theophilus Ademola Oluwaferanmi]]></dc:creator>
            <pubDate>Sun, 01 Nov 2026 01:59:13 GMT</pubDate>
            <atom:updated>2026-11-01T01:59:13.026Z</atom:updated>
            <content:encoded><![CDATA[<p><em>From UK milk to Nigerian cement, what a chemistry-turned-engineering student learned about grouping, cleaning and visualising real-world trade data.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/833/1*8AnLV1GTmdZsAwz93fue1w.png" /></figure><p>Everything looks better in pictures! Oh, they do. This week was about becoming Michelangelo and painting with data.</p><p>Things don’t really make sense to me until I see an application of the concept or get involved in building it. That’s what happened in week 8 of my Dataraflow data science internship.</p><p>I am a Civil Engineering student and before that I studied Industrial Chemistry. If there is one thing both fields taught me, it’s that you don’t trust a number until you know where it came from and how it was measured. This week, I found that the same rule applies to data.</p><h4>So, what was the week about?</h4><p>It was a continuation of the lessons from week 6, sharpening the tools every aspiring data scientist ought to be fortified with. The core of the week was grouping and aggregating data, along with different ways of visualizing it. We worked with UK milk imports and exports, and then repeated the whole analysis on trade data I chose myself. That part got me properly excited!</p><p>This article covers what I learnt, what I solved, where I struggled, what I’d also tell those in the same position.</p><h4>What I learnt</h4><p>How do you find out how much milk the UK imports or exports each month? You need to look at the database that holds such information. So the very first thing I learnt was how to go to the right source for the right data and how to enter the parameters based on what I am working on.</p><p>Once I had the data, what next? The lessons from weeks 6 and 7 become useful here. Comtrade data is so large that just looking at it would be counter-productive. To begin to reassemble the data into smaller, brain-downloadable chunks, I split the table into groups, summarize each group and combine the results. In Pandas that means groupby() with aggregations like sum, mean, count, min, max.</p><p>The question changed from ‘what is in this row?’ to ‘How much milk did the UK import from each country?’</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/678/1*DppnrDYZABMjBBMUU-Inpw.png" /><figcaption>One line of code turns thousands of rows into one answer per country</figcaption></figure><p>But the UK milk trade was only a test run, an example. Its tables already existed in a database, already prepared for my use. The next step was the one where I had to enter the parameters myself.</p><p>Then the whole concept of data visualization. But first, what did I solve?</p><h4>What I solved</h4><p>This week’s take home was divided into two: Part A and part B. Part A had two layers. First, I worked through the tasks solutions on grouping and aggregating data. Then came a test project which was an analysis of UK milk imports and exports. My approach was always the same:</p><ul><li>Look at the data first, (head()) and check what each column means.</li><li>Decide the question before writing code, such as which partners trade the most milk with the UK or how trade changes month to month.</li><li>Group, aggregate, sort, and check the output makes sense. Do the totals look reasonable?</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/530/1*3J03VIjcAq_ZHwE2xAGADA.png" /><figcaption>Looking at the data first: what each column means, and what the first fiverows contain.</figcaption></figure><h4>Repeating it with my own data</h4><p>Then I repeated the analysis but this time with my own Comtrade data. I chose cement and I was pleasantly surprised to find that Nigeria both exported and imported it! In fact, it baffled me. How does a country import what it exports?</p><p>On the Comtrade website, you set filters like these :</p><ul><li>Type of product(Goods or services),</li><li>Frequency(Yearly or monthly),</li><li>Period(months and years),</li><li>Reporter(The country whose data you want),</li><li>Partner(World or All),</li><li>Trade flow(Export or Import denoted as ‘X’ or ‘M’ in recent publications) ,</li><li>Commodity codes(HS): Here, you write a two- or four- digit code that represents the grouping of a particular commodity.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zalnlr-oM5WlU2kxHfFDNg.png" /><figcaption>The filters on Comtrade: product type, frequency, period, reporter, partner, trade flow and commodity code.</figcaption></figure><p>After previewing, I saw a mix of data that made no sense to me looking at them row by row even though they had column names. Why? There seemed to be no form of order or pattern between the rows. 202308 could be followed straight away by 202301. It was also a shock to see M and X instead of import and export. Then I found I could reorder the table itself on the Comtrade website, simply by using the ascending or descending arrows.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1hntRRrJj6iToNywYMNhWQ.png" /><figcaption>My first look at the data: periods out of order, and M and X in place of import and export.</figcaption></figure><p>Source: Comtrade Plus (comtradeplus.un.org)</p><p>I downloaded it, cleaned it, and ran the same steps as I did for the UK milk import.</p><p>Then I went on to visually represent my findings. Visualization adds a whole new dimension to data: at a glance, you can see what the dataset is trying to tell you.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/769/1*VhdS2FXf4Hs9HRRSttGF4Q.png" /><figcaption>Most UK milk trade records fall between about $1,000 and $1 million, with a tail of very small values on the left.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/696/1*RuuS53AvZzsOJ2kD--mhjg.png" /><figcaption>Germany supplied most of Nigeria’s cement imports in 2023, and the top three partners account for most of the World total.</figcaption></figure><p>The histogram shows how UK milk trade values from January to May 2015 are spread out. I used a log scale because the values range from a few dollars to over ten million, and on a normal axis the small ones would have been invisible. Most trade records fall between roughly $1,000 and $1 million, with the tallest bars around $100,000 to $1 million. A smaller group of very low-value records trails off to the left, and a few very large ones appear on the far right.</p><p>The bar chart compares Nigeria’s three biggest cement import partners in 2023 with the World total. Germany alone supplied about $4.2 million out of roughly $6.8 million, which is about 60% of everything Nigeria imported. China and the United Arab Emirates came next at around 1 million and 0.8 million. Together, the top three account for roughly 90% of the total, so the rest of the world contributes very little. I didn’t expect one country to supply so much of Nigeria’s cement</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/619/1*aODOziHD7gqWdt4myj1f7w.png" /><figcaption>Nigeria’s cement exports in 2023 went mostly to neighbouring countries, led by Niger, and were far larger than its imports.</figcaption></figure><p>To see the other side of the puzzle, I made the same chart for exports. Nigeria’s cement exports in 2023 were worth about $54 million, compared with about $6.8 million in imports, which is roughly eight times more. The destinations are also completely different. Niger took the largest share, at about $24 million, followed by Togo and Cameroon at around $14 million and $13 million. Together, these three account for most of Nigeria’s cement exports.</p><p>So how does a country import what it exports? Looking at the two charts, Nigeria isn’t swapping the same cement back and forth. It sells mostly to nearby countries and buys from far-away ones.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/593/1*SCcmMyOTkmUPxXlPaeG-Xg.png" /><figcaption>UK milk imports cluster at higher trade values than exports, which have a second group of smaller records</figcaption></figure><p>To compare imports and exports for UK milk, I drew a KDE plot, which is a smoothed version of the histogram with one curve per trade flow. Imports peak at around $1 million, with one clear hump. Exports have two humps, one at a few thousand dollars and another at around $100,000 to $1 million. The imports curve sits further to the right, which suggests that UK milk imports in this period were mostly larger-value records, while exports included many more smaller ones.</p><h4>Challenges I faced and how I overcame them</h4><p>There were many but the biggest was the gap between the tidy data the programme provided and real-world data. Mine had different column names, gaps and formats. My own data simply didn’t behave.</p><p>Also, I did not have enough time to revise and go over the work I had done in previous weeks. Between my engineering workload and other personal engagements, I couldn’t set aside a proper revision session. Some of this week’s ideas are still not fully settled in my head.</p><p>To overcome these, I:</p><ul><li>checked before aggregating, looking for nulls, duplicates and wrong data types. The truth is that most of my wrong results came from there.</li><li>ran my analysis one step at a time so it didn’t get tangled in my head.</li></ul><p>I also tried hard this week to avoid a habit that has dogged me for much of this internship: procrastinating and starting late. I started earlier. I first tried to revise previous lessons, and when I felt time moving too fast, I began the work by making sure I understood exactly what we’d been asked to do.</p><h4>A word for other learners</h4><ul><li>Learn the why and the code will make more sense.</li><li>Explore your data first before you start any work.</li><li>Be honest about what you have not mastered yet. Why? Because knowing your gaps is the first step to closing them.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=628f15a03e8c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Week 7: The Art of Cleaning, Visualizing, Trasforming and Combining Data.]]></title>
            <link>https://medium.com/@theophilusadeleye/week-7-the-art-of-cleaning-visualizing-trasforming-and-combining-data-5a329e5529f7?source=rss-fadf440416cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/5a329e5529f7</guid>
            <category><![CDATA[beginner]]></category>
            <category><![CDATA[civil-engineering]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[data-x-civil-eng]]></category>
            <dc:creator><![CDATA[Adeleye, Theophilus Ademola Oluwaferanmi]]></dc:creator>
            <pubDate>Sun, 25 Oct 2026 21:41:38 GMT</pubDate>
            <atom:updated>2026-10-25T21:41:38.607Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*4eRvHJzUnGJ7aEAM" /><figcaption>Photo by <a href="https://unsplash.com/@drwmrk?utm_source=medium&amp;utm_medium=referral">Andrew Stutesman</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Data, raw data. It is like buying meat from the butcher. That meat must be cleaned, but the next preparation processes depends not only on the type of meat, but also on the final use of the meat. Will it be Dry-Heat methods like barbecuing, grilling, roasting and baking, broiling or searing? or Frying methods like Sauteing, Deep-Frying? or Moist-Heat Methods? You get the idea.</p><p>The emphasis is on data in this case.</p><p>This week was about making that raw, messy data something that I can look at and trust the output, something that is meaningful to others when they see it, and then putting datasets together.</p><p>Well, this weeks course was partitioned into two - Part A: Cleaning and Visualizing Data, where I learnt how to clean column names, remove characters that were best referred to as junks, fix data types, handle missing values as well as use dates as an index; and, Part B: Transforming and Combining Data where I learnt how to write helper functions, join tables with pd.merge, use constants, and measure relationships with the spearman rank correlation and scatter plots. Thereafter, I practised the same ideas in SQL</p><p>Come join me on this journey of cooking this sumptuous data meal.</p><h4>Is Data available?</h4><p>This is a critical question. Before you go to your local meat seller, or butcher shop, of course you need to know if he has that particular meat available. If he doesn’t, you might have to visit another shop. It is the same with data. The 2014 London data was not ready to use. It was available (Lucky me!) but it wasn’t ready to use. It was downloaded as a CSV(Comma Separated Values) file. The column names had spaces(Does it matter? Yes it does. ‘ Events’ is not the same thing as ‘Events’. One has space before the E), one column was called WindDegrees&lt; /&gt; and its values ended in HTML, the dates were stored as text, and even the Events column had missing values.</p><p>Taking care of all these issues is what the cleaning part of A entails. I had to write this code to take care of the cleaning.</p><p>Before cleaning, this was how the table looked like:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kJ4yz0kKTQX4KZYEKU_QVA.png" /></figure><pre>london.columns = london.columns.str.strip()<br>london = london.rename(columns={&#39;WindDirDegrees&lt;br /&gt;&#39;: &#39;WindDirDegrees&#39;})<br>london[&#39;WindDirDegrees&#39;] = london[&#39;WindDirDegrees&#39;].str.rstrip(&#39;&lt;br /&gt;&#39;)<br>london[&#39;WindDirDegrees&#39;] = london[&#39;WindDirDegrees&#39;].astype(&#39;int64&#39;)<br>london[&#39;GMT&#39;] = pd.to_datetime(london[&#39;GMT&#39;])<br>london[&#39;Events&#39;] = london[&#39;Events&#39;].fillna(&#39;&#39;)</pre><p><strong>strip </strong>would strip it of rogue spaces, <strong>rename </strong>would help in taking care of a specific column name. <strong>rstrip </strong>would give the ability to change those names that have html tags which came to being when the dataset was exported from the Weather Underground and web site html line breaks were automatically added to each line in the file and which the read_csv() has interpreted as part of the column name and its values. The datatype of the columns had to be checked too. Some datatype would not let you operate on them.</p><p>After cleaning, the table would look like this</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LVCGl3fongCPNRHFWQMnaA.png" /></figure><p>GMT was datetime64 and WindDirDegrees was int64, so I could compare and filter them properly.</p><p>Setting the date as the index let me select by date range. I used it to plot humidity for spring.</p><h3>Planning a Summer Holiday?</h3><h4>The Holiday Weather from Project A</h4><p>You want to plan a summer holiday for you and your family? For a two weeks stretch? Of course, you would target a cluster of weeks where you will have zero rainfall.</p><p>I used a Weather Underground file for London in 2023 that I downloaded myself. Creating the CSV file had to be done manually by taking the data month by month.</p><p>In the Northern hemisphere, June, July and August are the summer months. so I treated June, July and August as summer and scored every possible 14-day stretch by its average temperature.</p><pre>weather[&#39;Date&#39;] = pd.to_datetime(weather[&#39;Date&#39;])<br>weather.index = weather[&#39;Date&#39;]<br>summer = weather.loc[datetime(2023,6,1):datetime(2023,8,31)]<br><br>rows = []<br>for i in range(len(summer) - 13):<br>    window = summer.iloc[i:i+14]<br>    rows.append([window.index[0], window.index[-1],<br>                 window[&#39;Temperature (C) Avg&#39;].mean(),<br>                 window[&#39;Precipitation Total (mm)&#39;].sum()])<br><br>windows = pd.DataFrame(columns=[&#39;Start&#39;, &#39;End&#39;, &#39;Mean temp (C)&#39;, &#39;Total rain (mm)&#39;], data=rows)<br>windows.sort_values(&#39;Mean temp (C)&#39;, ascending=False).head(3)</pre><p>The output was:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*XB64SR_PVaP_BNBVNkxokw.png" /></figure><p>My choice for the best two-week holiday is 12 to 25 June. There was one problem. The rainfall column in my download was 0.0 on every day of the year, which is not realistic, so I based my decision on temperature only.</p><h3>Part B: Transforming and combining data.</h3><h4>Combining Datasets.</h4><p>This part was one that made me laugh. It is built on the whole concept of JOIN- left, right, inner, outer.</p><p>I began by building small tables for GDP, life expectancy and population, and then joined them. Left, right, outer and inner joins keep different rows. The inner join keeps only the countries that are in both table.</p><pre>gdpVsLife = pd.merge(gdp, life, on=COUNTRY, how=&#39;inner&#39;)</pre><pre>for how in [&#39;left&#39;, &#39;right&#39;, &#39;outer&#39;, &#39;inner&#39;]:<br>    print(how)<br>    print(pd.merge(gdpVsLife, population, on=COUNTRY, how=how))</pre><p>I also rewrote repeated strings as constants, so I only need to change a column name in one place:</p><pre>COUNTRY = &#39;Country name&#39;<br>GDP_GBP = &#39;GDP (£m)&#39;<br>LIFE = &#39;Life expectancy (years)&#39;<br>POP = &#39;Population (millions)&#39;</pre><p>But then, things switched up. We had to put on the stove and let the fires lick the bottom the pot holding our meat and necessary seasoning.</p><h4>Real World Bank data</h4><p>Here, I worked with using the full data from the World Bank. It is possible to directly download data from the World Bank, for a particular time period and indicator. The first 34 rows of the World Bank tables are groups of countries, not single countries. I removed them and dropped the missing values before merging.</p><p>What was this task about? The analysis was simply to check whether there is any correlation between the total gross domestic product (GDP) of a country in 2013 and the life expectancy of people born in that country in 2013.</p><p>Then I had to take another dimension; Does population size relate to life expectancy? For my own project, I wanted to try a different question: <strong>do countries with bigger populations have shorter or longer life expectancy?</strong></p><p>I used two World Bank indicators for 2013 from the cohort database (schema week08):</p><ul><li>SP.POP.TOTL: total population</li><li>SP.DYN.LE00.IN: life expectancy at birth</li></ul><p>So I began by merging.</p><pre>popVsLife = pd.merge(popClean, lifeClean, on=COUNTRY, how=&#39;inner&#39;)<br>popVsLife.head()</pre><p>Then finding the correlation</p><pre>(correlation, pValue) = spearmanr(popVsLife[POP_INDICATOR], popVsLife[LIFE])<br>print(&#39;The correlation is&#39;, correlation)<br>if pValue &lt; 0.05:<br>    print(&#39;It is statistically significant.&#39;)<br>else:<br>    print(&#39;It is not statistically significant.&#39;)</pre><p>And displaying this result with a scatter plot.</p><pre>ax = popVsLife.plot(x=POP_INDICATOR, y=LIFE, kind=&#39;scatter&#39;, grid=True, logx=True, figsize=(10, 4))<br>ax.set_title(&#39;Population vs life expectancy, 2013&#39;)<br>ax.set_xlabel(&#39;Population (log scale)&#39;)<br>ax.set_ylabel(&#39;Life expectancy (years)&#39;)The coefficient was close to 0 and slightly negative, so there is no meaningful rank relationship between a country’s population and its life expectancy. Big countries aren’t reliably longer- or shorter-lived than small ones. </pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/842/1*RuzXVTQiTctRPochqKHxEw.png" /></figure><h3>Challenges and how I overcame them</h3><p>Column that looked clean but was not: My rainfall column had no missing values, but it was 0.0 every day. I checked the whole year and compared it with the London 2014 data, which has real rain, and then I based my holiday choice on temperature only.</p><p>Variables disappearing: After I restarted the kernel, I got NameError for run_query, weather and summer. I learnt to rerun the cells in order from the top. Along with a stuck database connection (PendingRollbackError). Restarting the kernel fixed it.</p><p>I had a missing practice file: The SQL setup failed with ModuleNotFoundError. I listed my db folder with os.listdir, saw that the file was missing, and created practice.py from the learning materials.</p><h3>A Word?</h3><p>Make sure that after loading data, you check dtypes, isnull().sum()and describe() before analysing anything.</p><p>Also, correlation is not causation, even a strong one.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5a329e5529f7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[193 Days of Rain or Snow, 0.00 Rainfall: A Lesson in Questioning Data]]></title>
            <link>https://medium.com/@theophilusadeleye/193-days-of-rain-or-snow-0-00-rainfall-a-lesson-in-questioning-data-768e5f0cd4d4?source=rss-fadf440416cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/768e5f0cd4d4</guid>
            <category><![CDATA[civil-engineering]]></category>
            <category><![CDATA[sql]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[data-science-training]]></category>
            <category><![CDATA[machine-learning-ai]]></category>
            <dc:creator><![CDATA[Adeleye, Theophilus Ademola Oluwaferanmi]]></dc:creator>
            <pubDate>Mon, 19 Oct 2026 15:04:37 GMT</pubDate>
            <atom:updated>2026-10-19T15:04:37.241Z</atom:updated>
            <content:encoded><![CDATA[<h4>Ten SQL Tasks, 120 quiz questions, and the patience it takes to learn Data Science as a beginner.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zb1xzzBOhihu_5Ge" /><figcaption>Photo by <a href="https://unsplash.com/@dkomow?utm_source=medium&amp;utm_medium=referral">Daniil Komov</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><blockquote>Do your best. But like it! Like what you do; then you will do your best — Katherine Johnson</blockquote><p>For a beginner like me, learning data science is not about how much code you can write or submit as final work, sometimes, it is about how long you can sit with a problem that seems insurmountable, how willing you are to question a result that otherwise looks perfect when it is far past your bedtime, and whether you can find the courage to try again… and again…and again when the code breaks. It is about being smart and planning ahead even when many external factors beyond your control are directly and indirectly influencing you.</p><p>I should have known there was going to be fire on the mountain when the founder announced we would enjoy the week’s task but of course, things are clearer in retrospect. This week, I worked with intermediate SQL, a database that was live, Numpy and Pandas. This week stretched me to yielding point that I sincerely questioned if I was making progress, if I was capable.</p><h3>Tool and Concepts Behind the Story</h3><p>The story begins with what we call a database, which is a structured collection of data stored electronically on a computer system, and a dataset which is an organized collection of related data points designed for analysis, research or machine learning.</p><p>Every aspiring data scientist is a disciple of the art of querying datasets. SQL (Structured Query Language), something looking like a shortened form of squirrel, is a standardized programming language used to access, manage, and manipulate data stored in a relational database management system. Like a squirrel assessing nuts it had stored before winter, to get information from such a management system, , one must do what is termed “Query”.</p><h4>Querying for Beginners: Understanding the Order of a SQL Query.</h4><p>Writing basic SQL queries was never the goal. To solve real world problem is the ultimate goal. To do this, one must be able to investigate real datasets and answer more complex questions. Working with intermediate SQL techniques required I learnt to filter with WHERE, to handle missing values in these datasets with NULL checks, using CASE statements for conditional logic, grouping records with GROUP BY and HAVING. I learnt to join Tables, identify unmatched records with anti-joins, and using Common Table Expressions (CTEs) to organize queries into more manageable steps.</p><p>One thing I began to understand while working with SQL is that the order in which we write a query is not the same as the order in which SQL logically processes it.</p><p>Take my first Week 6 task: <strong>finding the number of columns in each table in the database schema.</strong></p><pre>SELECT TABLE_NAME, COUNT(*) AS column_count<br>FROM INFORMATION_SCHEMA.COLUMNS<br>WHERE TABLE_SCHEMA = &#39;week06&#39;<br>GROUP BY TABLE_NAME<br>ORDER BY TABLE_NAME;</pre><p>Here is how the query works, in terms of logical execution.</p><ol><li>FROM identifies the source of the information. In this case, it is INFORMATION_SCHEMA.COLUMNS that provides metadata about the database columns.</li><li>WHERE filters the records, keeping only columns belonging to the week06 schema.</li><li>GROUP BY groups the records by table name so that each table can be counted separately.</li><li>COUNT(*) counts the column records in each group.</li><li>SELECT specifies the information to display: each table&#39;s name and its column count.</li><li>ORDER BY sorts the results by table name.</li></ol><p>Notice that we write SELECT first, but SQL&#39;s logical processing order begins with FROM. Understanding this distinction helps explain why clauses such as WHERE and GROUP BY work the way they do. SQL becomes easier when you understand not just what each clause does, but also how the clauses work together.</p><h4>Access and Assessments</h4><p>Like I said earlier, the SQL assessment covered quite a lot. It was particularly helpful because the questions required I applied several concepts rather than simply reproducing examples from a lesson.</p><ul><li>It began with exploring an unfamiliar database with INFORMATION_SCHEMA to understand its tables and columns before writing queries. Who starts building a house without checking carrying out a survey?</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/956/1*R-iwlsCUjgV0QC7MKNIo5A.png" /></figure><ul><li>Then proceeding to filtering data and handling NULL values correctly. NULL does not necessarily mean that the data does not exist, instead it means it is “unknown”. A comparison like =NULL never matches and I had to learn how to us IS NULL instead. &lt;&gt; silently drops NULL rows. This I used in Task 2 to count London’s days with no recorded event.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1006/1*amHNe5W57Aw8Ywmd2w6Avg.png" /></figure><ul><li>Using CASE expressions and conditional statements to count records that meet particular conditions.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Zj0rETiAYN3nmXJwNycgDQ.png" /></figure><ul><li>Applying GROUP BY and HAVING(Yes, the word “HAVING”) to summarize data and filter groups.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/693/1*wK91XURyIXHh6zyCxM0jaQ.png" /></figure><ul><li>Using joins and anti-joins. These young men, who sound like john and aunty john allow the one who queries to combine datasets and identify records that do not have a matching entry in another table.</li><li>Using UNION ALL to stack compatible datasets, the key word being “compatible”. Common Table Expressions also known as CTEs to organize complex queries and window functions to rank records and calculate running totals.</li></ul><h4>Why Does It Matter? The Weather Analysis</h4><p>All these are still rather abstract to a considerable degree. Appreciation comes from seeing a problem being solved with these things. The first SQL question had to do with a dataset related to weather analysis. Why? What do I need it for?</p><p>I was working with a year’s worth of weather records from six cities- Beijing, Brasilia, Cape Town, Delhi, London, and Moscow. But before comparing temperatures or rainfall, I needed to establish whether the data was complete, consistent, and reliable enough to support those comparison. But why? You may ask.</p><p>Imagine trying to investigate flooding in a Nigerian city using rainfall records from several locations. If some months are missing, some measurements are recorded differently, or there is a suspicious value, the resulting analysis could give an inaccurate picture of rainfall patterns.</p><p>So this is the thing- Learning to inspect database structures, identify missing values and compare related records is therefore not just a technical exercise. It is part of making sure that the evidence we use is worth trusting.</p><p>Do you see it now?</p><p>This was quite glaring when I examined Moscow’s weather data. The rainfall column showed 0.00 throughout the year, yet the events column recorded rain or snow on 193 days. If I had accepted the data without further investigation, it could have led me to draw a very misleading conclusion about the city’s weather.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/881/1*4c0XYRMNcIS90vjeinw9RQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*vnoEN_q2QE_p4flpcGPPJA.png" /></figure><p>The first picture, at the bottom shows where Moscow’s rainfall column showed 0.00.</p><p>Further querying it gave the second picture and showed that it wasn’t the case.</p><p>With London’s data, I learnt why missing values require careful treatment. I had to distinguish between records where the weather description was missing, records that mentioned rain and records where precipitation was greater than zero despite a missing event description. They are very different conditions and treating them as though they meant the same thing would have distorted the truth.</p><p>This has practical implications for my civil engineering domain. Suppose I wanted to study whether rainfall patterns were associated with flooding in particular communities. A missing weather description would not automatically mean that no rain occurred, just as an unrecorded flood incident would not mean that flooding never happened.</p><p>Another thing was combining information from different sources. In my work with Tubercolosis statistics from The World Health Organization and the Population data from the World Bank, I needed to bring the two datasets together to calculate death per 100,000 people.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/871/1*i96F5yJCRfx-2Tb5xMIkpg.png" /></figure><p>Why? Raw counts alone can make comparisons between populations of different sizes misleading. It did not stop at combining the tables with writing join, Country names such as “Viet Nam” and “Vietnam” did not match exactly. This is another reason why data from different sources must be standardized before it can be combined reliably.</p><p>Bringing it to my country, in a Nigerian infrastructure study, similar inconsistencies could arise when records from different agencies refer to the same community using different spellings or administrative labels. If those differences are not investigated, relevant records could be left out of the final analysis.</p><h4>Challenges Along the Way</h4><p>Thankfully, I had no problem with the set up. Problem started with the environment I was running the code in. I had kept the file path I used while moving my materials from an old workstation to a new one. The environments were clashing. I could run with one environment, only to discover it was the wrong one because the data I needed wasn’t there. But that was nothing compared to when I began with the learning materials and spent a whooping 30 hours trying to absorb in totality the content of the material. The quizzes at the end of the materials were now timed and required me to read between the lines. Still these were soft challenges.</p><p>While working on the SQL assessment, I kept having a PendingRollbackError. I restarted the notebook kernel and tried to run the set up, only to be hit with another error: TDS server connection failed.</p><p>It was frustrating because I needed the live connection to the database to execute my queries and check my answers. Investigating the issue revealed that my notebook and Python interpreter were running from different-looking project folders. I also discovered that the database helper and .env file were located inside the Week 6 folder rather than in the parent project directory I had initially checked.</p><p>That was a useful reminder that folder structures can be confusing, particularly when a project contains nested directories with similar names. A file can appear to be missing simply because I am checking the wrong location. Even after locating the database helper and confirming that the .env file existed, the SQL Server connection error persisted. At that point, I had not established the exact cause of the failure.</p><p>It was a lesson in patience and in being careful about the conclusions I draw when troubleshooting. Finding a possible source of a problem is not the same as proving that it caused the problem.</p><h3>How I approached the challenges and what I would do differently</h3><p>I tried to approach the problems by checking one thing at a time rather than immediately changing several parts of my setup. For the database issue, I checked the notebook’s working directory, the Python executable, the location of the database helper, and the presence of the .env file. These checks helped me understand the folder structure and narrow down what still needed investigation.</p><p>For the assessment itself, I want to become more deliberate about checking my work.</p><ul><li>Examining the structure of the data before writing a complicated query.</li><li>Verifying column names rather than assuming them.</li><li>Checking the output against the precise requirements of the question.</li><li>Keeping better notes of the errors I encounter and the steps I take to resolve them.</li></ul><h3>Advice for others learning data science</h3><p>If there is anything I would share with someone going through a similar learning process, it would be these lessons:</p><ul><li><strong>Understand the question before writing the code.</strong> Knowing the syntax is useful, but knowing what the result should represent is even more important.</li><li><strong>Explore your data before analyzing it.</strong> Check the number of rows, the column names, missing values, and whether related measurements make sense together.</li><li><strong>Pay attention to small details.</strong> A leading space in a column name, an incorrect filter, or integer division can change the result of an otherwise reasonable query.</li><li><strong>Do not assume every error means your code is wrong.</strong> Sometimes the issue is the environment, the file path, the connection, or the configuration.</li><li><strong>Do not change everything at once when troubleshooting.</strong> Check one possible cause at a time so you can understand what actually makes a difference.</li><li><strong>Keep practising what you have already learnt.</strong> Revisiting earlier concepts is not going backwards; it is part of becoming more independent.</li></ul><p>I am learning that progress is not always about how quickly I can write a query. Sometimes, it is about knowing how to investigate a problem, recognize an unreliable result, or explain why a particular approach makes sense.</p><h3>The workload, perseverance, and being at Dataraflow</h3><p>Beyond the technical work, this week also made me reflect on the experience of learning within Dataraflow.</p><p>There are moments when the workload feels heavy. Between learning materials, practical exercises, quizzes, assessments, and troubleshooting, there is a lot to process. Sometimes, you want to move quickly, but you also know that rushing through everything without understanding it properly will only create more difficulties later.</p><p>I am learning that perseverance is not necessarily about doing everything perfectly or never getting tired. It is about continuing to take useful steps, being willing to ask questions, and returning to a difficult problem with a better understanding of what needs to be done.</p><p>One quote that captures this feeling for me is:</p><p><em>“You do not have to carry the whole workload in one moment; you only have to keep taking the next step.”</em></p><p>What makes this journey even more meaningful is being part of an organization where the founders are intentional about our growth. I appreciate being in an environment where I am challenged to learn, practice, and develop skills that can be useful beyond the immediate assignment.</p><p>For me, this is not just about completing a data science program. With my civil engineering background, I want to understand how data can help investigate practical problems in my domain, from infrastructure performance to flooding and drainage challenges in Nigeria. I am still building the foundation, but I am beginning to see how the things I learn can eventually contribute to the questions I care about.</p><p>I am grateful for the opportunity to learn in an environment where development is intentional, even when the process itself is demanding.</p><h3>Final thoughts</h3><p>Week 6 taught me that learning data science involves more than understanding new functions or writing queries that run successfully. It requires me to think critically, check assumptions, work through unfamiliar problems, and recognize when I need to investigate further.</p><p>It also reminded me that I am still learning, and that there is value in being honest about what I understand, what I have completed, and what still needs work. The difficult moments did not make the week meaningless. If anything, they reminded me why practical experience matters.</p><p>I know there will be more demanding weeks ahead, but I am learning to take the process one step at a time. And I am grateful to be on this journey at Dataraflow, learning not only how to work with data, but also how to persevere, think more carefully, and grow into the kind of data professional I hope to become.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=768e5f0cd4d4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Dataraflow Week 4 and 5: Five possibilities, one choice.]]></title>
            <link>https://medium.com/@theophilusadeleye/dataraflow-week-4-and-5-five-possibilities-one-choice-29d10c27428f?source=rss-fadf440416cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/29d10c27428f</guid>
            <category><![CDATA[civil-engineering]]></category>
            <category><![CDATA[research]]></category>
            <category><![CDATA[beginner]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[data]]></category>
            <dc:creator><![CDATA[Adeleye, Theophilus Ademola Oluwaferanmi]]></dc:creator>
            <pubDate>Tue, 13 Oct 2026 07:03:09 GMT</pubDate>
            <atom:updated>2026-10-13T07:03:09.228Z</atom:updated>
            <content:encoded><![CDATA[<h4>Research topic, proposal and workflow</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*B5OpeK1NGKKBcjYp" /><figcaption>Photo by <a href="https://unsplash.com/@dandimmock?utm_source=medium&amp;utm_medium=referral">Dan Dimmock</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>If you think having limited options or choices is annoying, wait until you have a multitude of those options and then you understand why too many cooks can spoil the broth. I was cooking- yes, I was. When week 5’s notification dropped in my mail, my initial thought was “eh, just another week”. We had been given week 4 and 5 free from the usual drills of study materials and were asked to combine both weeks, find a suitable research topic, get it approved, write a research proposal and voila, our week is complete.</p><p>Coincidentally, it was the same week 4 and 5 I got to experience the role available power supply plays in the life of everyone using a laptop and the reason why tenacity is the base requirement for anyone who wants to succeed in life.</p><p>As you must have rightly guessed, I thought having no codes to write meant no work. How wrong I was, how wrong.</p><p>The issue was this: Finding a research topic sounds simple until you have to actually choose one. It is at that moment that every problem you have noticed seems like a possible research direction. I was walking through a street and noticed how silt, debris had blocked a sizeable portion of a large open-drain so much so that the volume of water coming out on the other side of the drain had reduced considerably. I thought “oh, this is a problem. Could drainage shape affect how these things come to be? Can I work on the most effective drainage shape to take care of issue like this?”</p><p>Oh, how little I knew! When I got back home and sat with my laptop and began to ask questions, it was then issues like the surrounding terrain, rainfall, maintenance practices began to pop up. Even availability of datasets to comfortably run such research. How do I take sample? How applicable is this research to my civil engineering field?</p><p>As a civil engineering student, the theoretical was not a problem for me. Classroom concepts such as water flow, hydraulic radius, strength of materials etc were things I had spent time learning and I can see clearly how they show up in everyday life especially in a country like Nigeria where infrastructure problems can affect people’s movement, businesses, homes and general quality of life.</p><p>Flooding has become a problem in my nation. Flood prediction for the year is not looking good either. You listen to the radio and you are scared for the possibilities. So imagine how excited I was that my mind was travelling to a direction like this. For the first time in my five years of studying civil engineering, I began to feel that I was capable of contributing meaningfully to my field and had begun to think like a professional and not like a student.</p><p>It became clear to me that there were many things I needed to discover.</p><p>So the very first lesson of this weeks 4 and 5- Being interested in a problem is not the same as solving it. But I am glad for one other thing: Learning data science with Dataraflow has changed the way I look at things around me. Definitely! Now I have what you would describe as a mentality shift! Say Glory somebody!</p><p>Now back to the topic. A topic can sound impressive, yet difficult to navigate. Another topic might seem ordinary at first but become valuable when you examine it closely. What really matters? Can that problem be defined properly? Is there enough evidence to study it? Can that work be completed with the time and resources available?</p><p>But reality hit fast. As beautiful and altruistic as my intentions were with respect to my proposed research topic, I had to let it go. I had even submitted it for review. But after attending meetings Dataraflow had in their wisdom and gracefulness organized for us, I had to tell my mentor “Please do not accept my research topic”</p><p>Who does that? You might be asking.</p><p>Well, it is at this point you will understand why the title of this article is five possibilities, one choice. I had to reach out to a lecturer I respected so much who was into research work. I had to carefully explain the situation to him. Mind you that time was already running out by now. People had already submitted week five’s articles and here I was still stuck in week 4. The desperation was real. My heart sank several inches.</p><p>I would end up with five topics and then the real work began. Which of these topics had a clearly defined gap. I found out that research was all about finding a gap and addressing it and not some phantom idea to try to do everything. I had to use every tool available to settle on a topic and when I did, I had to narrow it down, get the research gap, confirm it was something that I could bring a new dimension to.</p><p>Eventually, I submitted and it was approved.</p><h3>To Write a Proposal</h3><p>To write a proposal makes you confront what you do not know. Yes, believe me, it is mighty uncomfortable to plan a study in writing. When an idea exists only in your head, it is easy to assume that everything will fall in place but writing it down exposes the gaps. As I write this article, I am still working on the research proposal. Week six’ workload has dropped and I am yet to begin.</p><p>When writing a proposal, you may realise that your research topic is too broad, that your objectives do not quite match the problem, or that you have not carefully thought enough about where your data will come from.</p><p>The essential thing to note, is that these discoveries do not necessarily mean that the idea has failed, it is just part of developing it into something workable.</p><p>A useful lesson for week 5 is that choosing a topic and writing a proposal are not two completely separate tasks. In fact, the process of writing the proposal can reveal whether the topic is clear enough in the first place.</p><p>To help me, I had to use a mind mapping tool and to be sincere, sometimes I still feel like I do not know the full scope of what I am doing yet until I see the final result. I think that is just insecurity speaking.</p><h3>To The Next Phase</h3><p>These two weeks have made me sit up and begin to question what I am really doing here and what week 24 would look like for me. I want to be a civil engineer well grounded in data science, an analytical thinker and one with the ability to be in a place and contribute meaningfully by finding answers to problems.</p><p>Nigeria has many infrastructure and environmental challenges that deserve careful investigation but identifying a challenge is only the beginning. The work requires you to have a specific question, a realistic plan, evidence that is reliable and more importantly, conclusions that will not go beyond what the findings support. Oh, this is essential. Your conclusion must not go beyond your findings.</p><p>As I continue at Dataraflow, I hope I become better at turning observations from everyday life into research questions that can be investigated with discipline, that are supported by data and that are actually grounded in the realities of the environment around me.</p><h3>To The Reader</h3><p>To those reading, you can’t solve everything. Narrow that ambition down to something and work at it with a laser-like focus and see what comes out of it. Be careful though of having tunnel vision. In all, possibilities abound!</p><p>So long an article!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=29d10c27428f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond the Code]]></title>
            <link>https://medium.com/@theophilusadeleye/beyond-the-code-c50d17ab9a88?source=rss-fadf440416cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/c50d17ab9a88</guid>
            <category><![CDATA[data-science-training]]></category>
            <category><![CDATA[code-newbie]]></category>
            <category><![CDATA[python-programming]]></category>
            <category><![CDATA[civil-engineering]]></category>
            <category><![CDATA[data]]></category>
            <dc:creator><![CDATA[Adeleye, Theophilus Ademola Oluwaferanmi]]></dc:creator>
            <pubDate>Sat, 26 Sep 2026 20:34:29 GMT</pubDate>
            <atom:updated>2026-09-26T20:34:29.250Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>Connecting the dots between Python, databases, and version control to build real-world solutions.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*8qjlHNSbbp-DeYnk" /><figcaption>Photo by <a href="https://unsplash.com/@kvncnls?utm_source=medium&amp;utm_medium=referral">Kevin Canlas</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Week 3 came with another long list of Python topics: modules, dates, math, JSON, PIP, virtual environments, error handling, files, and an introduction to SQL.</p><p>All same same, but different different.</p><p>As days passed, I kept wondering, “Is there a connection between these topics?” By the end of the week, I started seeing the connection.</p><p>The good thing? I was no longer just learning how to make Python calculate something. <em>I was learning how to handle information, save it, retrieve it, and prepare programs for things that can go wrong.</em></p><h3>From using information to saving it</h3><p>One of the main things I learned this week was that <strong>Information doesn’t always have to disappear when the program ends. </strong>In fact, learning how to safely retrieve and manipulate informationit as well as datasets and databases felt like the core lesson of week 3.</p><p>That concept wasn’t glaringly obvious when I started, but it finally clicked when I began working on the JSON assignment.</p><p>I had to collect student records containing names, ages and grades, store them in a JSON file, and then load them back into Python.</p><p>The actual code wasn’t too long:</p><pre>with open(&quot;students.json&quot;, &quot;w&quot;) as file:<br>    json.dump(students, file)<br><br>with open(&quot;students.json&quot;, &quot;r&quot;) as file:<br>    loaded_students = json.load(file)</pre><p>But I initially had to understand what was happening here.</p><p>json.dump() takes Python data and puts it into a JSON file.</p><p>json.load() takes the JSON data from the file and brings it back into Python.</p><p>Far from being just another pair of functions to memorise, that distinction finally clicked for me. I was learning persistence, how information can be stored and brought back later.</p><p>I can already see how useful this could become in civil engineering. Site records, material inventories, project activities, inspection results or equipment information are all examples of data that could be collected, stored and retrieved rather than kept only in temporary variables or scattered records.</p><h3>Dates humbled me a little</h3><p>Dates gave me one of my more memorable errors this week.</p><p>I had written:</p><pre>datetime.strptime(first_date_time, &quot;%d/%m/%y&quot;)</pre><p>and Python responded with:</p><pre>ValueError: unconverted data remains: 25</pre><p>At first, the error looked more complicated than it actually was.</p><p>The problem was that I was entering a four-digit year such as 25/09/2025, while %y tells Python to expect a <strong>two-digit year</strong>.</p><p>I needed %Y:</p><pre>datetime.strptime(first_date_time, &quot;%d/%m/%Y&quot;)</pre><p>It was a small mistake, but it reinforced something I’ve been learning throughout this program: the computer is usually doing exactly what I told it to do, not what I intended it to do.</p><p>And dates aren’t only useful for birthday calculators. In my field as a civil engineering student, I can see they can be applied to civil engineering work — tracking project deadlines, calculating durations between activities, monitoring inspection dates, or keeping records of when site activities were completed.</p><h3>The while loop had its own lesson to teach</h3><p>One of the assessments asked me to calculate the GCD(Greatest Common Divisor) of two numbers.</p><p>I could have simply used:</p><pre>math.gcd(num1, num2)</pre><p>It works. But I wanted to understand what was happening underneath.</p><p>That led me to the Euclidean algorithm(which I did not know before, despite being a former Mathematics teacher) and, of course, to the while loop.</p><p>Initially, I wasn’t completely comfortable with while loops but the basic idea was:</p><blockquote><em>Keep doing something as long as a condition remains true.</em></blockquote><p>For GCD, the algorithm repeatedly calculates the remainder:</p><pre>while num2 != 0:<br>    remainder = num1 % num2<br>    num1 = num2<br>    num2 = remainder</pre><p>When num2 eventually becomes zero, num1 is the GCD.</p><p>Using the concepts of Euclidean algorithm was one of my favourite moments this week because I wasn’t just using Python’s built-in answer anymore. I was beginning to understand the <strong>algorithm behind the answer</strong>.</p><h3>A LinkedIn post changed how I looked at projects</h3><p>Somewhere in the middle of all this learning, I came across a LinkedIn post from someone who had combined some of the Week 3 concepts into a project.</p><p>What caught my attention wasn’t necessarily the project itself. It was the <strong>combination</strong>.</p><p>The person had taken things we were learning separately and turned them into something that looked like an actual application. It made me realise that these topics don’t have to remain isolated exercises that I complete one after another.</p><p>I started thinking:</p><p><strong>What could I build by combining what I already know?</strong></p><p>It coloured the way I looked at the remaining tasks for the rest of the week.</p><p>Instead of seeing JSON, dates, math, error handling, files and SQL as individual boxes to tick, I began to see them as pieces that could eventually fit together.</p><p>It made me think about my own background in civil engineering.</p><p>For example, instead of simply building another calculator, I could eventually create something that stores project information, calculates quantities, tracks dates, handles incorrect input and saves records for later use.</p><p>I’m still at the beginner stage, but the LinkedIn post gave me a different perspective: <strong>learning a technology is one thing; learning how to combine what you know into something useful is another.</strong></p><h3>Then came virtual environments</h3><p>PIP and virtual environments were another area where I initially saw commands before seeing the bigger picture.</p><p>Commands like:</p><pre>python -m venv venv<br>pip install numpy<br>pip list</pre><p>can easily become something you memorise and forget.</p><p>But the purpose is more important.</p><p>A virtual environment gives a project its own isolated space for packages and dependencies.</p><p>That matters because different projects may require different packages — or even different versions of the same package.</p><p>We also started getting introduced to MySQL.</p><p>Even something as basic as VARCHAR became useful because it showed me how databases think about information.</p><p>VARCHAR stores text with a variable length.</p><p>That might sound small, but it made me start thinking beyond Python variables and files.</p><p>If I eventually build something for civil engineering, I could have information about projects, materials, costs, dates, suppliers and activities that needs to be stored and retrieved systematically.</p><p>That’s where databases start becoming more than just another technology to learn.</p><h3>Git and GitHub: learning how to manage the work</h3><p>Our Zoom session with Mr. Henry introduced us to Git and GitHub.</p><p>This was different from the Python exercises because it wasn’t mainly about writing a new program.</p><p>It was about what happens <strong>around the code: </strong>tracking work, managing projects, collaborating, and eventually having a place where the things I build can be documented.</p><p>That also connects with the personal-brand conversation from an earlier week.</p><p>If I’m going to grow in data science, I don’t want my learning to exist only on my laptop.</p><p>I want there to be evidence of the things I’m learning and building.</p><h3>What Week 3 really taught me</h3><p>Looking back, I don’t think Week 3 was simply about completing seven Python topics and a series of assessments.</p><p>It was about moving one step beyond basic programming.</p><p>I learned that data can be stored and retrieved.</p><p>I learned that errors can be anticipated and handled.</p><p>I learned that built-in functions often have algorithms underneath them.</p><p>I learned that projects need isolated environments and dependencies.</p><p>I started thinking about how dates, stored data and databases could eventually solve practical problems in civil engineering.</p><p>And, thanks to that LinkedIn post, I started looking differently at the relationship between <strong>learning and building</strong>.</p><p>Our Zoom session with Mr. Henry also gave me my first steps toward version control with Git and GitHub.</p><p>I’m still very much learning. I try to use the leverage of having other interns around me but I try not to get sucked into the bottomless hole of feeling like I am never good enough.</p><p>I still make mistakes that sometimes take me longer than they should to understand.</p><p>But there is one question I find myself asking more often now:</p><p><strong>What is happening behind this line of code?</strong></p><p>And perhaps another question is beginning to follow it:</p><p><strong>What can I build with what I now know?</strong></p><p>For me, those might be the most important lessons from Week 3.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c50d17ab9a88" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I underestimated my data science workload. Then Friday came.]]></title>
            <link>https://medium.com/@theophilusadeleye/i-underestimated-my-data-science-workload-then-friday-came-c09c0c11efa4?source=rss-fadf440416cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/c09c0c11efa4</guid>
            <category><![CDATA[python-programming]]></category>
            <category><![CDATA[civil-engineering]]></category>
            <category><![CDATA[beginner]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[learning-to-code]]></category>
            <dc:creator><![CDATA[Adeleye, Theophilus Ademola Oluwaferanmi]]></dc:creator>
            <pubDate>Sun, 20 Sep 2026 12:39:26 GMT</pubDate>
            <atom:updated>2026-09-20T12:39:26.201Z</atom:updated>
            <content:encoded><![CDATA[<h4>Week 2: Lessons in Python, Object-oriented Programming, and the cost of Procrastination.</h4><figure><img alt="An picture of a shelf holding books in a library" src="https://cdn-images-1.medium.com/max/1024/0*g2PyDDuxWwGMcAjR" /><figcaption>Photo by <a href="https://unsplash.com/@emilwidlund?utm_source=medium&amp;utm_medium=referral">Emil Widlund</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><blockquote>Classes, inheritance, polymorphism, iterators, data structures, a library management system and the consequences of leaving too much work until the end of the week.</blockquote><h4>I Thought I Was Just Learning Python</h4><p>Last week, part of the things I learnt was that division by zero could make my code crash and how to prepare in advance for it. This week, which is the second week of my journey in data science, I learnt that procrastination too could make me crash and how to prepare for it too.</p><p>This week, I continued my journey into Python, and it was jarring to realize I wasn’t just learning to code, I was learning how to structure ideas. Because, why else am I doing class, then inheritance, iteration, scope?</p><p>Week 2 data science study dropped on Sunday last week. I opened the folder on Monday and saw Object Oriented Programming(OOP), then inheritance, iterators, polymorphism, scope, and I thought- “Well, that doesn’t look much”. I scrolled to the take home assignment and it looked relatively compact.</p><p>I had erroneously assumed it would be as simple as week one and so could leave answering the questions till late in the week. Oh my! I was so wrong. By Friday, I had not written a single code. I was still jumping from one tutorial to another, trying to understand concepts, all the while seeing my fellow interns making submissions as early as Wednesday.</p><p>3pm Saturday, 19th September, I was still building a library management system while on a general zoom meeting with the founder and other interns. In addition to this, the group I had joined for the founder’s challenge had started rolling out plans for the next week.</p><p>Here’s how the journey has been so far.</p><h4>Objects in Object-Oriented Programming: Just Terminologies or Practical Things?</h4><blockquote>In Python, I am beginning to learn I am not a bystander, I am a creator of things, of objects and I must think as a manufacturer does.</blockquote><p>The basis of object oriented programming is to visualize everything around me as an object. My chair, my table, my phone are objects. They each have properties that describe them and the things they can do.</p><p>A fan is an object, and it has certain properties which may or may not be related to the behavior of the fan. For example,</p><p>Properties of a fan: Brand, speed, color, size.</p><p>Behavior of a fan: Turn on, turn off, increase speed, decrease speed.</p><p>In python, properties are just simply referred to as ‘attributes’ and Behaviors as ‘methods’</p><p>Think of a car too, with attributes such as brand, model, color, and Methods such as start, accelerate, brake, stop etc.</p><p>These things are beyond terminologies, they are the way codes are structured.</p><p>In a manufacturing yard, where I will manufacture my objects, there must be a blueprint to follow. You don’t just go into production without a blueprint.</p><p>In Civil Engineering, there is a building plan. Several houses(Objects) can be built from that plan but the plan is not the physical house I will live in. It is only a blueprint. A spatial representation.</p><p>Python call these blueprints- CLASS</p><h4><strong>Classes and Objects</strong></h4><p>A class is like a blueprint, an object is something created from that blueprint. It therefore means that every object created in python must have a class. Logical, right?</p><p>Attributes will describe what that object has, while methods describe what the object can do.</p><p>Object Oriented programming(OOP) is a way of designing programs by viewing them as objects that have data(What they know) and behaviors(What they can do).</p><p>One thing that helped OOP finally click for me was looking away from my laptop. The room around me was already full of objects. My fan, my books, my phone has information about it and things it can do. OOP was beginning to feel less like a programming concept and more like a way of describing the world in a code.</p><p>A class of cars should contain objects such as Volvo, Tesla, BMW, Volkswagen, Mercedes Benz, Toyota, etc.</p><h4>Doing the Tasks</h4><p>The OOP tasks were relatively simple. They were the ones that made me erroneously assume that all the task were too simple to be taken serious. For example, this task required I write a code that will return a person’s name and his age.</p><pre>class Person:<br>  def __init__(self, name, age): <br>    self.name = name<br>    self.age = age<br>  def __str__(self):<br>    return f&quot;{self.name}({self.age})&quot;<br>p1 = Person(&#39;John&#39;, 36)<br>print(p1)</pre><p>By the end of tasks one, I knew the format for creating class.</p><ul><li>write class, followed by the class name</li><li>initialize the class by using def, __init__()</li><li>fill the init bracket first with the self parameter, then any other parameters I am given. In the code above, it is name and age, which define the object, p1 that I created.</li><li>use self.whatever parameters you have been given.</li></ul><p>It was this part that did not make sense at first. Why write self. I understood it as something that referred to the current object. But understanding the idea is different from actually getting comfortable with it.</p><p>When I write, self.name = name, the name on the left belongs to the current object, while the name on the right is the value being passed into the method. It sounds simple when explained, but it becomes a different story when you are the one writing the code</p><h4>Then Came Inheritance</h4><p>Once class started making sense, I mean being able to write it, inheritance introduced another layer. As the name implies, it is the part of OOP that caters to transfer of properties from parent class to children class. Or base class to sub class.</p><p>Tasks like creating a class BankAccount and a SavingsAccount where the savings account inherited the properties and behaviors from the bank account.</p><pre>class BankAccount:<br>    def __init__(self, owner, balance):<br>        self.owner = owner<br>        self.balance = balance<br><br>class SavingsAccount(BankAccount):<br>    def __init__(self, owner, balance, interest_rate):<br>        super().__init__(owner, balance)<br>        self.interest_rate = interest_rate</pre><p>This was where super() became important.</p><p>Instead of rewriting everything the parent class already knows how to do, the child class can reuse the functionality. I encountered the same idea again with creating the ElectricCar sub class from the vehicle class.</p><p>The deeper I went, the more I realized that inheritance wasn’t just about making code shorter, it was about representing relationships between things.</p><p>And it even became more glaring when I got to Polymorphism.</p><h4>Same Method. Different Behavior: Welcome Polymorphism.</h4><p>Polymorphism was one of those concepts that sounds very complicated than it actually is.</p><p>For example,</p><pre>class Dog:          #Create a Dog class<br>    def sound(self):      #define the method sound()<br>        return &quot;Woof!&quot;<br><br>class Cat:<br>    def sound(self):<br>        return &quot;Meow!&quot;<br><br>dog = Dog()        #create a dog object<br>cat = Cat()         #create a cat object<br><br>print(dog.sound()) #call the method, sound() on dog and print the result.<br>print(cat.sound())  #call the method sound() on cat and print the result</pre><p>Both objects have a method called sound(), but they behave differently when that method is called. For dog, it will return ‘Woof!’ and for cat, it will return ‘Meow!’</p><p>So, same method but different behavior. That was when this started to feel like I was beginning to model the real world.</p><h4>Then Abstract Classes</h4><p>Abstract classes helped me understand the difference between defining a requirement and defining an implementation. By using ABC and @abstractmethod, I could say that every shape must have an area() method, while leaving each specific shape to decide how that area is calculated.</p><pre>from abc import ABC, abstractmethod<br><br>class Shape(ABC):<br><br>    @abstractmethod<br>    def area(self):<br>        pass<br><br>class Rectangle(Shape):<br>    def __init__(self, length, breadth):<br>        self.length = length<br>        self.breadth = breadth<br><br>    def area(self):<br>        return self.length * self. breadth<br><br>class Circle(Shape):<br>    def __init__(self, radius):<br>        self.radius = radius<br><br>    def area(self):<br>        return 3.142 * (self.radius ** 2)<br><br>shapes = [Rectangle(3, 4), Circle(5), Rectangle(5, 6), Circle(3)]<br><br>for shape in shapes:<br>    print(shape.area())</pre><h4>And Then I Started Making Mistakes</h4><p>This was probably the most useful part of the week. Because I made mistakes. I made so many mistakes they got imprinted in my eyes.</p><p>At one point, I wrote:</p><pre>count =+ 1</pre><p>When I meant:</p><pre>count += 1</pre><p>They look almost identical. They are not. The first means: count = count + 1</p><p>The second, count =+ 1 effectively assigns positive one to count.</p><p>I also initially placed a break outside a loop. I had a typo where I wrote EvenNmbers instead of EvenNumbers. Some other times, I forgot the self in a method.</p><p>These mistakes were small but they taught me that:</p><blockquote>This is less about knowing what the correct code looks like and more about being able to understand why your code is wrong, especially from the error the system gives as feedback</blockquote><p>This is a skill I still have a lot to develop.</p><h4>The part I didn’t expect: Iterators</h4><p>I built a countdown iterator. No, please don’t shout yay! yet.</p><pre>class Countdown:<br><br>    def __init__(self, n):<br>         self.n = n<br>         <br>    def __iter__(self):<br>        return self<br><br>    def __next__(self):<br>        if self.n &lt;= 0:<br>            raise StopIteration<br><br><br>        x = self.n<br>        self.n -= 1<br>        return x<br><br>mycountdown = Countdown(5)<br><br>for number in mycountdown: <br>    print(number)</pre><p>Before iterators, I had used for loops without really thinking about what was happening underneath. Now, I i have learnt that Python uses protocol behind the scenes.</p><p>__iter__() tells Python how to obtain the iterator.</p><p>__next__() tells it how to get the next value, and</p><p>StopIteration tells it when there are no more values.</p><p>I went further and built an iterator for even numbers, a file-line reader, and even a Fibonacci iterator. The interesting thing was that these weren’t just exercises in syntax. They forced me to think about <strong>how data moves through a program</strong>.</p><h3>Scope Had Me Looking Twice</h3><p>Then came scope. Then LEGB(Local Enclosing Global Built-in). I learned about global and nonlocal but one example that particularly caught me off guard was the one about accidentally using a name like len inside a function.</p><p>If Python sees an assignment to len anywhere inside that function, it treats len as a local variable throughout that function. That means something as innocent-looking as:</p><pre>len(numbers)</pre><p>can cause trouble if I later assign to len inside the same function.</p><p>It was a reminder that even names matter in programming. Sometimes the problem isn’t the logic you’re trying to write. Sometimes it’s the name you chose.</p><h4>Then I Had to Build Something Bigger</h4><p>Eventually, the individual concepts had to come together. I worked on a Library Management System.</p><p>This was where the week stopped feeling like a collection of disconnected Python lessons. I created a Book class. Then I extended it into EBook and PrintedBook.</p><p>I created members and different types of members. I introduced borrowing and returning books. I used class variables to keep track of total borrowed books. I used private attributes. I implemented searching. I even made the library and members iterable.</p><p>Suddenly, the concepts I had been learning separately were interacting inside one system.</p><p>And that was probably the biggest lesson of the week:</p><blockquote>Knowing a concept individually is one thing. Knowing when and how to combine it is another.</blockquote><p>I wrote a complete Library Management System using these concepts. You can view the full code on my</p><pre>class Book: <br>    def __init__(self, title, author, year):<br>        self.title = title<br>        self.author = author<br>        self.year = year<br>        self.is_available = True<br><br><br>    def get_description(self):<br>        return f&quot;{self.title} by {self.author} ({self.year})&quot;<br><br><br>class EBook(Book):<br>    def __init__(self, title, author, year, file_size, format):<br>        super().__init__(title, author, year)<br>        self.file_size = file_size<br>        self.format = format<br><br><br>    def get_description(self):<br>        return (<br>            f&quot;EBook: {self.title} by {self.author}&quot;<br>            f&quot;({self.year}), {self.file_size}MB, {self.format}&quot;   <br>        )<br><br>class PrintedBook(Book):<br>    def __init__(self, title, author, year, weight, dimension):<br>        super().__init__(title, author, year)<br>        self.weight = weight<br>        self.dimension = dimension<br><br>    def get_description(self):<br>        return (<br>            f&quot;Printed Book: {self.title} by {self.author}  &quot;<br>            f&quot;({self.year}), {self.weight}kg, {self.dimension}&quot;<br>        )<br><br><br></pre><p>I wrote a complete Library Management System using these concepts. You can view the full code and my entire coding journey in data science on Github.</p><p><a href="https://github.com/Adetanol/My-Week-2-journey--Library-Management-System/tree/main">GitHub - Adetanol/My-Week-2-journey--Library-Management-System</a></p><h3>But Here’s What Made This Week Difficult</h3><p>The technical content wasn’t the only challenge. My biggest mistake was thinking I had more time than I actually did. I kept looking at the week as if the work would somehow become lighter because I hadn’t started it yet. It didn’t.</p><p>The assignments were still there. The videos were still there. The concepts still needed to be understood and as the week progressed, everything I had postponed began arriving at the same time.</p><p>Then Friday came.</p><p>And suddenly, the workload didn’t look like a list of small tasks anymore. It looked like one large problem. That experience made something very obvious to me:</p><blockquote>Procrastination doesn’t remove work. It only changes when you experience the pressure of doing it.</blockquote><p>And in a field like data science, where one concept can depend on another, postponing work can make things even harder. If I don’t properly understand one concept today, tomorrow’s lesson doesn’t necessarily wait for me to catch up.</p><p>Between Friday, Saturday and Sunday, I spent a total of about 30 hours on my computer. It is not funny.</p><h3>Data Structures and Algorithms Put Things Into Perspective</h3><p>One of the materials for the week also introduced me to data structures and algorithms. At a basic level, data structures are ways of organizing data, while algorithms are the procedures used to work with that data.</p><p>But what interested me was the bigger idea. The way we structure information affects what we can do with it. That is not entirely foreign to me. Coming from the civil engineering domain, I already understand that how a problem is structured can influence how effectively it can be solved.</p><p>A road network, for example, isn’t just a collection of roads. There are relationships between locations, routes, traffic patterns and constraints. Flooding isn’t simply “water on a road.” There are rainfall patterns, drainage systems, elevations, soil conditions, land use and other variables interacting with one another.</p><p>Construction delays aren’t just delays. They can involve materials, labour, cost, scheduling, weather and logistics. Data science gives me another way of thinking about these kinds of problems.</p><p>I’m beginning to see that the domain doesn’t disappear when I learn technology.</p><h3>Taking the Lesson Outside the Laptop</h3><p>One thing I found interesting this week was realizing that I could apply some of these programming ideas to ordinary things around me.</p><p>Take the books around my room.</p><p>I could think of a Book as an object with properties such as:</p><ul><li>title</li><li>author</li><li>year</li><li>format</li></ul><p>It could also have behaviours such as:</p><ul><li>display its description</li><li>be borrowed</li><li>be returned</li></ul><p>An eBook could inherit from the general idea of a book while adding things like file size and format. A physical book could have weight and dimensions.</p><p>I could now see that the Library Management System wasn’t just an exercise, it was a way of representing something I could actually see around me. This led me to a question I want to keep asking as I continue learning:</p><blockquote>What are the objects around me? What information do they contain? What can they do? How are they related? And how could I represent all of this with data?</blockquote><p>That question feels more valuable to me than simply memorizing another Python keyword.</p><h3>Beyond the Code: Building a Personal Brand</h3><p>We also had a session with the founder of the program about building our personal brands. It was an interesting addition to a week that had been heavily focused on technical learning. Because learning something and being able to communicate that learning are two different skills.</p><p>I can spend hours learning Python, data structures or algorithms but if I cannot explain what I am learning, document my progress, or communicate the value of what I can do, part of that journey remains invisible.</p><p>For me, this is especially interesting because my background is in the civil engineering domain, while I’m deliberately expanding into data science, programming and technology.</p><p>Documenting the journey isn’t just about saying, <em>“I learned Python this week.” </em>It’s about creating a record of how my thinking is changing. What I struggled with. What I built. What I misunderstood. What I eventually understood and perhaps, over time, how these different areas of my experience begin to connect.</p><p>Build solid connections too. It helps. So I have decided to take social media seriously, my group members seriously, even those who are my subscribers and the readers of this article.</p><h3>What Friday Taught Me</h3><p>If there is one lesson I’m taking from this week that has nothing to do with Python syntax, it is this:</p><blockquote><strong>Don’t confuse having time with having unlimited time.</strong></blockquote><p>At the beginning of the week, the workload looked manageable. By the end of the week, the same workload felt much heavier because I had allowed it to accumulate. Ironically, the pressure didn’t make the learning better. It made the learning harder.</p><p>I don’t think the solution is to become someone who never procrastinates again. That would probably be an unrealistic promise. The more useful lesson is to recognize the cost earlier. Start the assignment when it is given. Watch the video before it becomes urgent. Try the code before the deadline forces the attempt. Give difficult concepts enough space to actually sink in.</p><p>Why? Sometimes the hardest part of learning isn’t the concept itself. It’s arriving at the concept tired, rushed and already thinking about everything else you haven’t done.</p><p>Ask questions too, go online.</p><h3>I Am Still Learning How to Learn</h3><p>Week 2 gave me classes, inheritance, polymorphism, abstraction, iterators, scope, data structures and algorithms. Underneath all of that, it gave me something else.</p><p>It showed me how much learning depends on <strong>consistency</strong>. I underestimated the workload.</p><p>Then Friday came.</p><p>Friday made the consequences impossible to ignore. But I’m glad I experienced it now, while I’m still learning how I learn.</p><p>The goal isn’t simply to finish each week’s material. I want to understand it. I want to be able to build with it. I want to recognize where it can become useful outside the classroom.</p><p>Eventually, I want to be able to take what I’m learning in data science and use it to ask better questions within my own domain. I’m not learning data science to abandon my domain. I’m learning another way to understand and solve problems within it.</p><p>If this week taught me anything, it’s that the learning itself deserves to be treated like the work it is, not something I can keep pushing to Friday and expect to magically become easier.</p><h4>Don’t Just Run Codes, Write it.</h4><p>I must commend the recent decision by dataraflow to take our learning materials from Github and place it within the portal so that we were forced to write codes in order to see the result it produced. On VS code, all I needed to do was press run but as from next week, I will be writing all of those codes myself. I believe it will help me as I have learned that writing codes makes it stick better.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c09c0c11efa4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My First Week in Learning Data Science: If curiosity killed the Cat, it will tame the Python.]]></title>
            <link>https://medium.com/@theophilusadeleye/my-first-week-in-learning-data-science-if-curiosity-killed-the-cat-it-will-tame-the-python-000a60fee294?source=rss-fadf440416cd------2</link>
            <guid isPermaLink="false">https://medium.com/p/000a60fee294</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[beginner]]></category>
            <category><![CDATA[data]]></category>
            <dc:creator><![CDATA[Adeleye, Theophilus Ademola Oluwaferanmi]]></dc:creator>
            <pubDate>Tue, 08 Sep 2026 15:06:31 GMT</pubDate>
            <atom:updated>2026-09-08T15:06:31.492Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*0qmIdkg13f5fMzhW" /><figcaption>Photo by <a href="https://unsplash.com/@blender_person?utm_source=medium&amp;utm_medium=referral">wesfly</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Table of content</p><ol><li>Introduction: Dipping my toes in the data flow.</li><li>Unmasking the digital masquerade: What is data science?</li><li>A crash course in Python: Crashing (and learning) through week one.</li><li>Challenges I faced: Thinking like a computer.</li><li>A word of advice: Embrace P.U.S.H.</li></ol><blockquote>As a data scientist, one must be curious and have the ability to create business value from data. A data scientist must also have skill set that includes Mathematics, Technology, Programming and business acumen.</blockquote><h4>Introduction: Dipping my toes in the flow called Data</h4><p>I have dabbled in data science one way or the other. It wasn’t difficult to come in contact with. It was like the friend of a friend you were bound to cross path with in the general room. A lecturer introduced me to it, for the sake of preparing us for a world ruled by data. At that time, I never understood why. I just accepted because every new idea looked fancy. But my relationship with data science even then was like bathing in a cold stream during December. I would put a part of my body, just a toe maybe, for a few hours and return several months later.</p><p>I decided to take it seriously this year and dedicated my time to a program but it was difficult. I couldn’t recall sessions I had finished, when I got to the project execution stage, it seemed I had learned nothing. I resorted to writing codes and whole lessons in a notebook. The result? Lessons that should take 10 minutes could take as long as 1 hour. I got frustrated, blame it on “I am not a natural coder” and went on holidays again. Until a friend shared an opportunity to learn Data science with Dataraflow’s 24-weeks programme, lasting for roughly six month.</p><p>I got admitted into the programme as part of the cohort 2 for 2026 and then it dawned on me as to what I had really signed up for.</p><p>But why did I not give up on data science? Considering how lackadaisical I had been about it. Why did I keep going back for opportunities related to studying it?</p><p>I have always felt that if my country would transition from being onlookers in the field of Civil Engineering, then data must play a critical role. From structural analysis to strength of materials and hydrological surveys, besides the calculations is a whole world of data churned out from these processes and undoubtedly would require Data mining, machine learning, domain knowledge, big data to fully harness the potentials. These things were clearly related to data science.</p><p>I knew this and I wanted this. I wanted to position myself for the future too, to be among those whose insight would help create the country we have always envisioned but the ability to do that lies in data science. Furthermore, I wanted to give back to the society, my country, to prepare the next generation for a more glorious future.</p><h4>Unmasking the Digital Masquerade: What is data science?</h4><blockquote>“Data science is an inter-disciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from many structured and unstructured data”.</blockquote><p>It is related to the mining of data, big data and machine learning. It is a</p><blockquote>“concept that unify statistics, data analysis, machine learning, domain knowledge and their related methods”</blockquote><p>The idea is not to bore you. If one is to do data science, one must be duly armed with the tools for data science. A writer must have a means to write, be it the traditional note and pen or an electronic device. Likewise, Data science requires programming language tools such as Python and R.</p><p>Why? Not only is Python and R open sourced and Python is a general purpose programming language, and R is for statistical analysis, they have machine learning libraries, visualization, data manipulation and analysis package. Now, these things are what every data scientist must have. Then add software tools like Anaconda, KoBoToolbox and SPSS</p><p>To data science, I needed to marry Machine Learning, which is “a branch of artificial intelligence that enables computer to learn and improve from experience without being explicitly programmed. Basically, an ML system uses data to recognize patterns, make decisions and improve its performance over time”</p><p>In simple analogy,</p><blockquote>Just like humans learn from examples and experiences, machines learn from data.</blockquote><p>It is through the power of data and machine learning that platforms like Netflix, Amazon, Facebook, Youtube and others have gotten to the height they have.</p><h4>A Crash Course in Python: Crashing(and Learning) Through Week One</h4><p>It is just week one and apart from installation and general introduction, we are to do a crash course in Python. Did I crash? Yes, I did. Did I learn? Yes, I learned and I am still learning. It was not easy transitioning from a laid-back approach to learning, to one that was structured and required me to be out of my comfort zone, weekly evaluations and study, with the ever lingering axe of termination from the program just floating over my head.</p><p>I learnt about the basic concepts of Python as a programming language, about data types like lists, tuple, dictionaries, strings, numeric data types like floats, integers. I have learnt about the for loops and while loops, when they are used, why they are used.</p><p>In the process of all this, I have learnt the use of Github and git, their push and pull, repositories, VS code and other necessary tools, whose use I would never have learnt(Or would have done so at a much later time) if I did not have this opportunity.</p><p>I have learnt that the best way to learn anything is to learn the basics and understand why, where and when those basics are used. I did not understand this at first, and began by using brute force to go through each learning requirement until I got to the “Take Home” assignment part. It was then my previous failures in other data science studies came back to me. I needed to know the basics, to develop skill sets entrenched in subjects like Mathematics.</p><p>To do is better than to read a multitude of pages. Python requires doing.</p><h4>Challenges I faced: Thinking Like a Computer.</h4><p>I have never thought myself intelligent enough to learn to code. I have always thought I was too rigid a learner to be flexible enough to learn a new language but I had to learn to think the way a computer thinks.</p><p>In dealing with python, I had to learn to stop treating calculations done by my computer the way I treat calculations in my head. 2 plus 3 is automatically 5 in my head, but I can’t just write 2 plus 3 in a computer. I have to tell the computer, create a space where a user can enter any number. From there comes the issue of creating variables, input function, even integers.</p><p>Check out this example:</p><p>Write a program that can perform addition, subtraction, multiplication and division based on users input. Of course, those are the rudimentary operations of a calculator. True but I am just a learner and I had to learn.</p><p>I needed to start by creating a variable called num1 to take in values.</p><p>Why, you may ask? Well, there must be an interface that asks the user to input their number. To add 2 and 3, I must be able to type in 2 and let the computer know I have typed in 2.</p><p>So I have num1 = float(input(“Enter the first number”))</p><p>The num1 is the variable name, that stores whatever I input, so it can be called later. It is like naming a human being.</p><p>What about float? It is to take care of the fact that the inputted numbers might have decimal points. There is no assumption with computers. All instructions must be expressly written.</p><p>Why is there a bracket? Oh, that word float defines the input and acts as a function and the other one, input, is also a function and functions defined whatever is in the bracket in front of them.</p><p>Why is Enter the first number in double quotation marks? Because it is a string. A string is a name for words and sentences in python.</p><p>So we have-</p><p>num1 = float(input(“Enter the first number”))</p><p>num2 = float(input(“Enter the second number”))</p><p>What do I do with the both numbers I have gotten? Don’t forget you have to perform those operations on them.</p><p>Operator = input(“Enter an operator (+ — * ?)”)</p><p>Well, after I have entered 2 and 3 and enter +, what do I do then? I have to instruct the computer what to do if a certain operator is entered. I have to use the If loop</p><p>if operator == “+”:</p><p>result = num1 + num2</p><p>print(round(result, 3))</p><p>Human translation- If the person enters +, create a temporary variable called result, where it will add num1 and num2 as it was earlier defined.</p><p>The print statement is always used in Python to display an output on the screen.</p><pre>operator = input(&quot;Enter an operator (+ - * /)&quot;)<br>num1 = float(input(&quot;Enter the first number&quot;))<br>num2 = float(input(&quot;Enter the second number&quot;))<br><br>if operator == &quot;+&quot;:<br>  result = num1 + num2<br>  print(round(result, 3))<br>elif operator == &quot;-&quot;:<br>  result = num1 - num2<br>  print(round(result, 3))<br>elif operator == &quot;*&quot;:<br>  result = num1 * num2<br>  print(round(result, 3))<br>elif operator == &quot;/&quot;:<br>    print(round(result, 3))<br>else:<br>  print(f&quot;{operator} is not valid&quot;)</pre><p>This might not be smooth especially to those who are well versed in the art of programming, because I am still learning but gradually, I am getting to understand concepts.</p><p>There are sometimes when a single letter or omission of one, such as a colon (:) might cause a bug. Sometimes, one has to be careful where they place certain function, like this code I wrote to find the second largest number in a list.</p><pre>my_list = []<br><br>while True:<br>  entry = (input(&quot;Enter the numbers here: (q to exit if done): &quot;))<br>  if entry.lower() == &quot;q&quot;:<br>      break<br>  else:<br>    my_list.append(int(entry))<br><br>print(f&quot;Your list is: {my_list}&quot;)<br><br>my_list.sort()<br>print(&quot;Second largest number is:&quot;, my_list[-2])</pre><p>The version that kept giving me a bug, happened to have ‘int’ in a different line.</p><pre>my_list = []<br><br>while True:<br>  entry = int(input(&quot;Enter the numbers here: (q to exit if done): &quot;))<br>  if entry.lower() == &quot;q&quot;:<br>      break<br>  else:<br>    my_list.append(entry)<br><br>print(f&quot;Your list is: {my_list}&quot;)<br><br>my_list.sort()<br>print(&quot;Second largest number is:&quot;, my_list[-2])</pre><p>It kept returning an error that int has no definition called lower. The difference in both codes is just where I placed int. In the first, I placed it in the line 7, while in the second, I placed it in line 3. Placing it in line three is a conflicting instruction as it tells the computer to make sure the inputs are integers. But if I am to enter q which is not an integer, an error then pops up. Yet I must enter q to terminate that program.</p><p>What did I do? Remove the ‘int’ from line three so that it accepts a form of input whether string or integers, and I placed it in line 7 where it the instruction is only executed after q must have been inputted.</p><p>Python is structural. It goes line by line and so sequence matters. But it does not mean all programmers will write the same code for a given program.</p><p>Oh, while editing this write up, I had to correct one to the codes I typed because Python codes are case-sensitive. If and if are not the same thing.</p><p>Another one taught me the basics of mathematics; Write a program that swaps two numbers</p><pre>num1 = int(input(&quot;Enter first number: &quot;))<br>num2 = int(input(&quot;Enter second number: &quot;))<br><br>print(&quot;First number: &quot;, num1)<br>print(&quot;Second number: &quot;, num2)<br><br>num1 = num1 + num2<br>num2 = num1 - num2<br>num1 = num1 - num2<br><br>print(&quot;After swapping: &quot;)<br>print(&quot;First number is:&quot;, num1)<br>print(&quot;Second numbeer is:&quot;, num2)</pre><p>The solution was simply to use the understanding of the properties of numbers- addition and substraction.</p><h4>A word of Advice: Embrace P.U.S.H</h4><p>The fundamentals are very essential. Never assume the basics don’t matter. Always learn to ask Why? Where? How? Because it is the application that will tell others if you really understood the basics. When you get to be give real life scenarios, how would you know where to start from.</p><p>Also, don’t hesitate to ask questions. Or ask for help from others.</p><p>Understand the role of AI as a tool and not your substitute student. As much as you just want to get over the exercises, you should remember that writing a program you cannot explain is tantamount to zero.</p><p>The truth is that I am not where I want to be, even in terms of knowledge. I still sit in front of my laptop and look at the wall while asking myself how to start writing a program. I am sure some more experienced people reading this would probably point out mistakes in my code. It is welcome. I am just starting and I am eager to learn.</p><p>This journey brought me to Medium as a platform, made me find new friends and learn to work with others, get frustrated and come back after eating, stare at the wall for hours and try not to copy another person’s work. I have learnt to fight the feeling of inadequacy too, because when I see the posts of others, I am tempted to clean all of the code I have written and give up… but I won’t. I am not where they are right now, but I will push(Python Until Something Happens or Press Until It Happens) and this is just week one.</p><p>I look forward to the end of the programme, with the mindset that all is possible to them who do the right thing at the right time and are not afraid to explore. I am certain that as basic as num1 plus num2 is, the understanding will come in useful as we dive deeper into data science, machine learning.</p><p>Is Python choking me? Yes, but will I let go of it? Nah!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=000a60fee294" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>