<?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 Balqisbankole on Medium]]></title>
        <description><![CDATA[Stories by Balqisbankole on Medium]]></description>
        <link>https://medium.com/@balqisbankole3?source=rss-52308e3b1ad9------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*updPZA7_JTxknl_P</url>
            <title>Stories by Balqisbankole on Medium</title>
            <link>https://medium.com/@balqisbankole3?source=rss-52308e3b1ad9------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 11:15:13 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@balqisbankole3/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[One Brain, Too Many Tabs Open]]></title>
            <link>https://medium.com/@balqisbankole3/one-brain-too-many-tabs-open-9aebe1457847?source=rss-52308e3b1ad9------2</link>
            <guid isPermaLink="false">https://medium.com/p/9aebe1457847</guid>
            <dc:creator><![CDATA[Balqisbankole]]></dc:creator>
            <pubDate>Wed, 04 Nov 2026 07:07:17 GMT</pubDate>
            <atom:updated>2026-11-04T07:07:17.004Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aexg-rT4_q0T5MNr_ulxyQ.png" /></figure><p><em>Week 8 of My Data Science Journey at DataraFlow</em></p><p>I really need to stop overestimating how much I can accomplish in 24 hours.</p><p>Because why did I look at my schedule last week and genuinely believe everything would work out?</p><p>I had clothes to sew, an event to attend in another state, a job interview to prepare for, and my weekly DataraFlow assignments.</p><p>Four different commitments, all demanding attention, and somehow, I was the only person available to handle them.</p><p>The plan was simple. Finish sewing early, travel for the event, prepare for my interview, and squeeze in some studying whenever I had free time.</p><p>The problem with that plan was the assumption that I would actually have free time.</p><h3><strong>When your to-do list has its own to-do list</strong></h3><p>You know that feeling when you’ve been busy all day, but the moment you sit down, you start remembering everything you haven’t done?</p><p>That was basically my week.</p><p>I’d spend hours sewing, finally sit down to rest, and remember that I hadn’t touched my assignment.</p><p>Then I’d open my laptop, stare at the notebook, and spend the first few minutes trying to remember what I was doing the last time I opened it.</p><p>Sometimes, I would read the same explanation three times and still not understand it. Not necessarily because it was difficult, but because my brain had already clocked out for the day.</p><p>There was also the familiar experience of telling myself, Let me just finish this one thing.</p><p>One thing would become another thing, and before I knew it, the day was practically over.</p><p>Meanwhile, the assignment deadline was getting closer.</p><h3>Apparently, grouping data isn’t just grouping data</h3><p>This week’s topic at DataraFlow was grouping and aggregating data.</p><p>At first glance, it looked manageable. We were learning how to use pandas to group records, calculate summaries, and compare results.</p><p>Nothing too frightening.</p><p>Then I started working on the actual tasks.</p><p>We were analysing international trade data, and I downloaded records of Nigeria’s imports of fresh apples, pears and quinces in 2023.</p><p>My first confusion came when I discovered that my dataset had 57 import records, but after filtering, only 45 remained.</p><p>I kept wondering what happened to the other 12.</p><p>Turns out, they were the World aggregate records. Removing them was necessary to avoid counting the same trade twice.</p><p>A simple explanation, but it took me longer than I’d like to admit to understand.</p><p>Then came the interesting part.</p><p>After grouping the data by trading partner, I discovered that South Africa accounted for over 94% of Nigeria’s recorded imports of those fruits.</p><p>I hadn’t expected such a huge difference.</p><p>It was one of those moments when the numbers stopped looking like random figures and started telling an actual story.</p><h3>SQL had other plans for me</h3><p>Just as I was getting comfortable with pandas, SQL decided to remind me that there was still plenty to learn.</p><p>We worked on conditional aggregation, GROUPING SETS, window functions, and ranking trading partners.</p><p>There were questions I understood immediately but couldn’t figure out how to translate into SQL.</p><p>And there were queries that looked perfectly reasonable until I ran them.</p><p>You know that little excitement you feel when your code finally runs without an error?</p><p>Only to look at the output and realise it isn’t what you expected.</p><p>So now you have a new problem.</p><p>I also discovered that understanding a solution while reading it doesn’t always mean you can write it yourself.</p><p>I’d look at an explanation and think, Oh, that makes sense.</p><p>Then I’d try to reproduce it without looking.</p><p>Suddenly, nothing made sense anymore.</p><p>At some point, I was switching between SQL queries, interview preparation, and sewing.</p><p>Imagine trying to remember the difference between GROUP BY and PARTITION BY while also thinking about whether you’ve finished the outfit you need for an event.</p><p>Not exactly the ideal learning environment.</p><h3>The part nobody puts in their weekly updates</h3><p>I think one of the hardest things about learning a new skill as an adult is that life doesn’t stop to give you time to learn.</p><p>There are days when you’re genuinely interested in what you’re studying, but you simply don’t have the energy for it.</p><p>You finally get an hour to yourself, open your laptop, and realise you’re too tired to concentrate.</p><p>Other times, you manage to understand something, feel proud of yourself, and then discover that the next task requires three new concepts you haven’t learned.</p><p>And somehow, everyone online seems to be completing projects, collecting certificates, and announcing their latest achievements.</p><p>Meanwhile, you’re just trying to remember whether you saved your notebook before closing VS Code.</p><p>Still, somewhere between all the stress, I noticed something encouraging.</p><p>I was starting to understand how to approach analytical questions.</p><p>Instead of immediately worrying about the code, I began thinking about what I needed to calculate, how the data should be grouped, and what the results would actually mean.</p><p>I’m also getting more comfortable moving between Python and SQL, even though I still have moments when I need to revisit things I’ve supposedly learned.</p><p>And when I finally created some visualisations with Matplotlib, seeing those figures on a chart made the effort feel worthwhile.</p><h3>Did I find the perfect balance?</h3><p>Absolutely not.</p><p>Some things took longer than planned. I underestimated how exhausted I’d be after travelling, and there were moments when I had to choose between getting some sleep and making progress on my assignment.</p><p>I also learned that scheduling five hours of studying doesn’t magically give you five hours of concentration.</p><p>Who knew?</p><p>But I’m beginning to appreciate the small improvements.</p><p>A few weeks ago, I would have struggled to understand some of the queries I worked on this week. Now, even when I can’t solve something immediately, I have a better idea of where to begin.</p><p>I’m not going to pretend that every difficult week comes with some profound life lesson.</p><p>Sometimes, a stressful week is simply a stressful week.</p><p>But I did learn something about myself.</p><p>I enjoy this enough to keep returning to it, even when my schedule makes absolutely no sense.</p><p>For Week 9, I’m hoping for fewer deadlines, fewer surprises, and a little more sleep.</p><p>Although, knowing me, I’ll probably look at my calendar and convince myself I can squeeze in one more thing.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9aebe1457847" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Think I’m Finally Beginning to Understand Research Statistics, and I’m Excited About It]]></title>
            <link>https://medium.com/@balqisbankole3/i-think-im-finally-beginning-to-understand-research-statistics-and-i-m-excited-about-it-08d9f5bdab01?source=rss-52308e3b1ad9------2</link>
            <guid isPermaLink="false">https://medium.com/p/08d9f5bdab01</guid>
            <dc:creator><![CDATA[Balqisbankole]]></dc:creator>
            <pubDate>Sun, 25 Oct 2026 03:38:59 GMT</pubDate>
            <atom:updated>2026-10-25T03:38:59.694Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sNg4BJEooWzdmA8nVCKocw.png" /></figure><p>Week 7 of my DataraFlow Data Science Internship gave me a new perspective on statistical analysis, programming, and some challenges I’ve been trying to overcome for years.</p><p>I think I’m finally beginning to crack the code in data science, or perhaps I should say research statistics.</p><p>I know that might sound a little dramatic, but I’m genuinely excited about what I’ve discovered this week. It’s one of those moments when you learn something and immediately start thinking about all the problems you could have solved if you’d known about it earlier.</p><p>The interesting thing is that this excitement has very little to do with learning a complicated new programming technique. It’s more about finally seeing how the things I’ve learned over the years can work together.</p><p>To explain why this means so much to me, I need to share a little about my background.</p><h3>Why Statistical Analysis Has Always Been Important to Me</h3><p>I recently completed my MSc in Plant Breeding and Genetics, and one thing I’ve always wanted is to be able to analyse my research data independently.</p><p>As a researcher, I understand the importance of experimental design, data collection, statistical analysis, and interpreting results. But I’ve never been comfortable with the idea of collecting data for months, sometimes under very demanding conditions, only to depend on someone else to carry out the analysis.</p><p>I wanted to understand every stage of the process and be confident enough to do it myself.</p><p>My interest in data analytics actually started before my master’s degree. I enrolled in a data analytics programme because I thought it would help me develop the skills I needed. However, as the programme progressed, I realised it focused mainly on business analytics.</p><p>I enjoyed learning about sales, business performance, and other applications, but I was looking for something more relevant to scientific research.</p><p>I wanted to know how to compare experimental treatments, test hypotheses, investigate relationships between traits, and use statistical methods to answer research questions.</p><p>That was one of the reasons I decided to pursue my master’s degree.</p><p>During the programme, I gained a better understanding of research methodology, experimental design, and the theoretical aspects of statistical analysis. However, I was disappointed that we weren’t really taught how to carry out these analyses independently using programming tools.</p><p>It was frustrating because I understood what needed to be done, but I still didn’t feel equipped to do it myself.</p><p>For a while, I wondered whether I would ever find the kind of training I was looking for.</p><h3>The Discovery That Made My Week</h3><p>This week at DataraFlow, we explored data transformation, combining datasets, visualisation, and statistical analysis using Python.</p><p>We worked with weather records, World Bank indicators, and relationships between different variables. As part of the lessons, I was introduced to statistical analysis using Python’s SciPy library.</p><p>And honestly, I couldn’t believe it.</p><p>You mean there’s a Python package that can perform statistical tests?</p><p>The same kinds of tests I’ve been wanting to carry out independently?</p><p>I know this is probably nothing new to people who have been programming for years, but discovering SciPy was genuinely exciting for me.</p><p>I started thinking about the statistical concepts I’d learned during my master’s degree and how I could begin applying them using Python.</p><p>Correlation, hypothesis testing, p-values, and statistical significance were concepts I already knew, at least theoretically. Now I was beginning to understand how to implement some of them.</p><p>Of course, knowing that a library can perform a statistical test doesn’t automatically mean you know when to use that test or how to interpret the results. There’s still a lot I need to learn.</p><p>But just knowing that I could develop these skills using Python made me incredibly happy.</p><p>I kept thinking about my previous research and how much I would have loved to have this knowledge then.</p><p>For the first time, I feel like I’m finding what I’ve been searching for all this while.</p><p>## You Really Don’t Know What You Don’t Know</p><p>Another thing that made this week particularly interesting was learning about logarithmic scales.</p><p>This might seem unrelated to my excitement about research statistics, but it reminded me how much a single concept can change your approach to a problem.</p><p>I’ve been working on a sales analysis project, and there was a particular chart I had been struggling with.</p><p>Some values were extremely large compared with others, and the chart looked strange. Most of the data points were concentrated in one small area, while a few extreme values stretched the scale.</p><p>I kept trying to figure out how to make the visualisation more meaningful.</p><p>Then I learned about logarithmic scales.</p><p>I immediately thought about that chart.</p><p>A logarithmic scale allows values to be displayed based on their ratios rather than equal numerical intervals. This can make it easier to examine patterns when values vary considerably in magnitude.</p><p>It doesn’t remove outliers, and it isn’t appropriate for every type of data, but it offers another way to explore distributions that are difficult to interpret on a linear scale.</p><p>I was excited because I could immediately see a possible application to something I was already working on.</p><p>It made me realise that sometimes, you can spend hours struggling with a problem simply because you haven’t encountered a particular concept or technique.</p><p>You don’t know what you don’t know, and that’s exactly why continuous learning is so important.</p><p>I haven’t forgotten that sales chart. I’m looking forward to revisiting it with this new understanding.</p><h3>Of Course, Python Had Other Plans</h3><p>As exciting as this week was, it also came with its fair share of frustration.</p><p>At one point, I couldn’t run my assessment because Python kept throwing errors.</p><p>First, it was a ModuleNotFoundError. After trying to resolve that, I encountered an ImportError, followed by a SyntaxError.</p><p>I spent a considerable amount of time checking folders, inspecting file paths, and trying to understand why Python couldn’t import a module that I thought was already available.</p><p>Eventually, I discovered that the file was initially empty. After adding the code, I encountered another problem because the formatting had been damaged.</p><p>Some imports had been combined incorrectly, and the indentation was completely wrong.</p><p>At that point, I wasn’t even working on the actual assignment anymore. I was trying to get the assessment environment to cooperate.</p><p>I also encountered a KeyError while trying to filter weather records by date. I had converted the date column into datetime format, but I hadn’t realised that this didn’t automatically make it the DataFrame’s index.</p><p>These were small things, but they took time to figure out.</p><p>One thing I’ve noticed about learning programming is that understanding a concept and successfully implementing it are sometimes two very different experiences.</p><p>You can know exactly what you want your code to do and still spend an hour trying to understand why it isn’t working.</p><p>I’m gradually learning to be more patient with that process. I’m also becoming more comfortable reading error messages, checking my assumptions, and troubleshooting problems systematically.</p><h3><strong>When Your Analysis Doesn’t Give You the Result You Expected</strong></h3><p>One of our projects this week involved choosing two World Bank indicators and investigating whether a relationship existed between them.</p><p>I decided to analyse inflation and unemployment using data from 2023.</p><p>After cleaning and combining the datasets, I had complete observations for 166 countries.</p><p>I used Spearman’s rank correlation to investigate whether countries with higher inflation also tended to have higher unemployment.</p><p>The correlation coefficient was approximately 0.053, indicating a negligible positive association. The result was also not statistically significant at the 5% level.</p><p>I found this particularly interesting because inflation and unemployment are often discussed together in economics.</p><p>However, my analysis did not show a statistically significant monotonic relationship between the two indicators across the countries studied.</p><p>That doesn’t mean inflation and unemployment are never related. My analysis compared countries in a single year, and relationships within individual countries over time could be quite different.</p><p>It reminded me that statistical analysis is not simply about calculating a correlation coefficient and deciding whether two variables are related.</p><p>You also need to understand the context, the limitations of your data, and what your results can reasonably tell you.</p><p>This is one of the things I enjoy about research. Sometimes, the results don’t support your expectations, but they give you something else to investigate.</p><h3>Bringing Everything Together</h3><p>We covered quite a lot this week.</p><p>I practised cleaning messy weather data, working with dates, handling missing values, transforming columns, and combining datasets using different types of joins.</p><p>I also worked on projects involving weather analysis, GDP and life expectancy, and inflation and unemployment.</p><p>In SQL, we practised cleaning text fields, filtering dates, joining tables, identifying missing records, and using window functions to calculate rainfall totals over consecutive days.</p><p>I completed all five SQL assessment tasks successfully, which was especially satisfying considering how much trouble I initially had getting the assessment to run.</p><p>But what stood out to me most wasn’t the number of tasks I completed.</p><p>It was how much more connected everything felt.</p><p>The statistical concepts I learned during my master’s degree are beginning to make more sense as I practise implementing them in Python. My previous experience with business analytics is also becoming useful in ways I hadn’t initially anticipated.</p><p>I’m beginning to see how these skills can complement one another, particularly as someone with a background in scientific research.</p><h3>A Different Kind of Progress</h3><p>I still have a lot to learn, especially when it comes to choosing appropriate statistical methods, understanding their assumptions, and applying them correctly to experimental research.</p><p>I’m also still learning Python, and I’m sure there will be many more errors to troubleshoot.</p><p>But I’m happy with the progress I’m making.</p><p>For years, I’ve wanted to become more independent in analysing scientific research data. This week made that goal feel more achievable.</p><p>It also reminded me that learning something new doesn’t always mean starting from scratch. Sometimes, you’re building on knowledge you already have, even when the connection isn’t immediately obvious.</p><p>I’m beginning to appreciate how much my background in research can contribute to my journey in data science, and how the technical skills I’m developing can strengthen my work as a researcher.</p><p>Perhaps that’s why this week’s learning felt so significant to me.</p><p>I haven’t mastered research statistics, and I certainly haven’t figured out everything about data science.</p><p>But I finally feel like I’m moving in the direction I’ve wanted to go for a long time.</p><p>And honestly, I’m really happy about that.</p><p>Have you ever discovered a tool or concept that made you realise there was a solution to something you’d been struggling with for ages? I’d love to hear about it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=08d9f5bdab01" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Seven Days, Too Many Errors, and a Deadline That Wasn’t Negotiating: My Week 6 Data Science…]]></title>
            <link>https://medium.com/@balqisbankole3/seven-days-too-many-errors-and-a-deadline-that-wasnt-negotiating-my-week-6-data-science-4caedf5c21d6?source=rss-52308e3b1ad9------2</link>
            <guid isPermaLink="false">https://medium.com/p/4caedf5c21d6</guid>
            <category><![CDATA[sql]]></category>
            <category><![CDATA[pandas]]></category>
            <category><![CDATA[sqlalchemy]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[numpy]]></category>
            <dc:creator><![CDATA[Balqisbankole]]></dc:creator>
            <pubDate>Sun, 18 Oct 2026 10:52:45 GMT</pubDate>
            <atom:updated>2026-10-18T10:52:45.227Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bFjgi0zTbOnaxWVEyO8ULQ.jpeg" /></figure><h3>Seven Days, Too Many Errors, and a Deadline That Wasn’t Negotiating: My Week 6 Data Science Experience</h3><p><em>What a week of NumPy, Pandas, intermediate SQL, timed quizzes, and a challenging assessment taught me about problem-solving, persistence, and the importance of questioning data.</em></p><p>There is a special kind of confidence that comes from seeing a familiar topic on your learning schedule.</p><p>That was me when I saw SQL.</p><p>Of all the tools I have been learning on my data analytics journey, SQL has always been one of my favourites. I genuinely enjoy writing queries, exploring tables, and watching rows of data turn into meaningful information.</p><p>So, when I saw Intermediate SQL on our Week 6 schedule at <strong>DataraFlow</strong>, I thought, <em>At least there is something here I am already comfortable with.</em></p><p>Well, SQL had other plans for me. 😂</p><p>By the time I started exploring Common Table Expressions (CTEs), window functions, and other unfamiliar concepts, I realised there was a lot more to SQL than I had imagined.</p><p>And SQL was only one part of the week.</p><p>There was NumPy. Pandas. Database connections. Practical exercises. A take-home assessment. Weather data analysis. Business-related SQL problems.</p><p>Then, somewhere in the middle of all that, there were monitored quizzes.</p><p><strong>Ten questions. Four minutes.</strong></p><p>As if the learning materials and weekly deliverable were not already enough to keep me busy.</p><p>And did I mention that failing to submit our weekly deliverable could mean being dropped from the programme?</p><p>So, whether you understood everything or not, the deadline was coming.</p><p>Honestly, at some point, I wasn’t even sure I would be able to meet up.</p><p>But strangely enough, despite how overwhelming everything felt, I actually enjoyed the experience.</p><p>Let me tell you how it went.</p><h3>Seven Days Is Not as Long as It Sounds</h3><p>You know that feeling when you open your learning portal, see several lessons waiting for you, and convince yourself that you will finish them quickly?</p><p>Then you open the first lesson.</p><p>You encounter something unfamiliar.</p><p>You decide to understand it properly before moving on.</p><p>Thirty minutes later, you are still trying to understand a concept you thought would take five minutes.</p><p>Meanwhile, the remaining lessons are patiently waiting for you.</p><p>That was the kind of week I was having.</p><p>One thing I quickly realised is that understanding a concept while reading an explanation and applying it independently are two very different experiences.</p><p>The first makes you feel intelligent.</p><p>The second tests that confidence. 😂</p><p>And with a strict deadline, I couldn’t afford to spend the entire week waiting until I understood everything perfectly.</p><p>I had to learn, practise, make mistakes, figure things out, and keep moving.</p><h3>NumPy, Pandas, and the Art of Making Sense of Data</h3><p>A significant part of the week focused on NumPy and Pandas.</p><p>With NumPy, I explored arrays, indexing, slicing, reshaping, joining, splitting, broadcasting, and statistical operations.</p><p>Some concepts seemed straightforward until I started working with multidimensional arrays.</p><p>And then there was axis.</p><p>Ah, axis. 😂</p><p>Just when I thought I understood it, another operation would make me stop and think again.</p><p>Reshaping arrays was another interesting experience. Apparently, you cannot rearrange 12 elements into a shape that requires 15.</p><p>Python will not manufacture extra elements just because you need them.</p><p>How inconsiderate.</p><p>But the more I experimented, the more I realised that visualising what I was doing was far more useful than memorising functions.</p><p>Pandas brought another set of practical lessons: selecting and filtering data, working with Series and DataFrames, converting data types, and handling missing or inconsistent values.</p><p>One exercise involved converting values containing both numbers and text into numerical data.</p><p>It reminded me that real-world datasets don’t always arrive neatly organised.</p><p>Sometimes, a column that should contain numbers has text sitting comfortably in the middle of it, minding its business.</p><p>And it becomes your responsibility to sort things out.</p><p><strong>My biggest lesson here was that understanding the structure and quality of your data should come before trying to analyse it.</strong></p><p>A calculation is only as useful as the data behind it.</p><h3>SQL, I Still Love You, But We Need to Talk</h3><p>Now, this was probably my favourite discovery of the week.</p><p>I thought I was already comfortable with SQL, but intermediate SQL introduced me to concepts I had barely explored.</p><p>CTEs, window functions, subqueries, conditional aggregation, different types of joins, and more.</p><p>Some were completely new to me. Others were things I had heard about but had never really used.</p><p>And I loved discovering them.</p><p>It felt like realising that a place you thought you knew well had several rooms you had never entered.</p><p>CTEs particularly caught my attention because they showed me how to organise complex queries into smaller, more understandable steps.</p><p>Instead of trying to solve everything in one long query, I could break the problem down and build the solution gradually.</p><p>Window functions were another exciting discovery, especially when I began exploring rankings and running totals.</p><p>But beyond the syntax, something else stood out to me.</p><p><strong>Sometimes, a problem isn’t necessarily too difficult. You’re just trying to solve too many parts of it at once.</strong></p><p>Breaking a complicated query into smaller steps made it easier to understand, test, and troubleshoot.</p><p>And honestly, that is a lesson I can apply far beyond SQL.</p><h3>When Connecting SQL to Python Became Its Own Assignment</h3><p>Another interesting part of the week involved connecting Python to our cohort’s Azure SQL database.</p><p>The idea was exciting. Instead of downloading datasets manually, I could retrieve information directly from the database and continue analysing it with Python.</p><p>Sounds straightforward, right?</p><p>Until database connection errors enter the conversation.</p><p>There were environment variables, configuration files, credentials, and several things that needed to work together.</p><p>At one point, I encountered transaction-related errors that refused to disappear.</p><p>You run your code.</p><p>Error.</p><p>You make an adjustment.</p><p>Another error.</p><p>And sometimes the traceback is longer than the code you originally wrote.</p><p>It is the kind of situation that makes you wonder whether the computer has something personal against you. 😂</p><p>But the experience taught me something useful about troubleshooting.</p><p>Randomly changing things and hoping something works is not always the best approach.</p><p>I needed to pay attention to error messages, check my environment, and understand whether the problem was coming from the connection, the query, or something else.</p><p>It also helped me appreciate how SQL and Python work together.</p><p>SQL can retrieve and organise the information I need, while Python provides additional tools for exploring, analysing, and visualising it.</p><h3>The Assessment: Where Everything Became Real</h3><p>The take-home assessment was where the week’s learning really came together.</p><p>One major task involved analysing <strong>2,190 daily weather records across six cities</strong>: Beijing, Brasília, Cape Town, Delhi, London, and Moscow.</p><p>I explored temperature patterns, precipitation, humidity, and extreme weather conditions.</p><p>Brasília recorded the highest average daily mean temperature at 22.9°C, while Moscow recorded the lowest at 6.0°C.</p><p>Beijing had the highest daily maximum temperature at 42°C, while Moscow recorded the lowest daily minimum at -26°C.</p><p>But one of the most interesting findings wasn’t about temperature.</p><p>It was about Moscow’s precipitation data.</p><p>According to the precipitation column, Moscow recorded <strong>zero precipitation throughout the entire year</strong>.</p><p>Not one wet day.</p><p>Yet another column contained records of rain and snow.</p><p>Now, that is suspicious.</p><p>Imagine presenting a report confidently stating that Moscow experienced no precipitation in 2014 simply because that’s what the numbers showed.</p><p>The query might be correct. The calculation might be correct. But the conclusion could still be misleading.</p><p>That was one of my biggest lessons from the project.</p><p><strong>Data doesn’t become reliable simply because it exists in a database.</strong></p><p>Sometimes, the most valuable thing an analyst can do is question a result that doesn’t make sense.</p><p>Before trusting an interesting finding, we need to ask: Does this result make sense? Is there other information that supports it? Could something be missing or incorrectly recorded?</p><p>These questions matter just as much as knowing how to write the code.</p><h3>When SQL Started Asking Business Questions</h3><p>The SQL assessment also involved customer churn, tuberculosis indicators, population data, milk exports, and loan approval information.</p><p>One task revealed that customers on two-year contracts had a much lower churn rate (27.1%) than customers on month-to-month contracts (57.6%).</p><p>That was interesting, but it also raised another question: what other factors might explain the difference?</p><p>A relationship in data doesn’t automatically prove that one factor caused another.</p><p>Another task involved joining WHO and World Bank datasets.</p><p>Sounds simple enough.</p><p>Except that <strong>28 countries failed to match</strong>.</p><p>Not necessarily because the countries were absent, but because different datasets can use different names for the same country.</p><p>Imagine trying to match two people in separate records when one uses a full name and the other uses a nickname.</p><p>You know they could be the same person, but the computer doesn’t automatically know that.</p><p>That’s essentially one of the challenges of combining datasets.</p><p>I also encountered situations where SQL calculations could produce misleading results because of data types, even when the query ran successfully.</p><p>And in a loan approval analysis, some results looked suspiciously perfect, with approval rates of 0% or 100%.</p><p>Those results needed further investigation before anyone could confidently use the data to support decisions or train a predictive model.</p><p>These tasks changed how I thought about analysis.</p><p><strong>A query running successfully doesn’t automatically mean the answer is correct. And a correct calculation doesn’t always mean the conclusion is reliable.</strong></p><p>The responsibility of a data analyst goes beyond producing numbers. It involves understanding what those numbers mean, questioning unusual patterns, and recognising when further investigation is necessary.</p><h3>Four Minutes, Ten Questions, and No Time for Overthinking</h3><p>As if all of that wasn’t enough, the learning portal also included monitored quizzes.</p><p>Ten questions in four minutes.</p><p>That’s an average of <strong>24 seconds per question</strong>.</p><p>Twenty-four seconds!</p><p>Sometimes, that is barely enough time to read the question, understand the options, and convince yourself that the answer you selected is correct.</p><p>You see a question and immediately recognise the topic.</p><p>You feel confident.</p><p>Then you read option A.</p><p>Reasonable.</p><p>Option B.</p><p>Also reasonable.</p><p>Option C.</p><p>Wait. Why does this one sound correct too?</p><p>And suddenly, a question you thought you understood has become a personal investigation.</p><p>Meanwhile, the timer has absolutely no interest in your internal debate. 😂</p><p>The quizzes made me realise the difference between recognising a concept and actually understanding it.</p><p>It’s one thing to read an explanation and think, <em>Oh, I know this.</em></p><p>It’s another to apply that knowledge when you’re presented with a question you haven’t seen before.</p><p>I learnt that testing myself, experimenting with examples, and explaining concepts in my own words are much more useful than simply reading through lessons and assuming I understand them.</p><h3>The Part Nobody Really Sees</h3><p>When people share their progress in data analytics, we usually see the final results.</p><p>A clean notebook.</p><p>A beautiful chart.</p><p>A successful query.</p><p>A completed assessment.</p><p>What we rarely see is everything that happened before those results appeared.</p><p>The query that returned an unexpected number of rows.</p><p>The missing bracket.</p><p>The code that worked earlier but suddenly started throwing errors.</p><p>The moment you realise you’ve been staring at the same problem for so long that you’re no longer sure what you’re trying to fix.</p><p>Or that funny experience of spending so much time debugging something, only to discover that the problem was a tiny detail.</p><p>Sometimes, the solution isn’t complicated.</p><p>You’re just too tired or too focused on one approach to notice it.</p><p>This week reminded me that learning to code requires patience, curiosity, and the willingness to be wrong several times before getting something right.</p><p>And sometimes, the most productive thing you can do is leave the screen briefly and return with a clearer mind.</p><p>Because staring harder at the same line of code doesn’t necessarily make it correct.</p><p>I wish it did, though. 😂</p><h3>Five Lessons I’m Taking Beyond Week 6</h3><p>Looking back, I realised that some of my biggest lessons had very little to do with memorising Python functions or SQL syntax.</p><p><strong>1. A successful result isn’t necessarily a correct result.</strong></p><p>Just because your code runs doesn’t mean your output is accurate. Check your calculations, question unexpected results, and make sure your conclusions are supported by the data.</p><p><strong>2. Don’t ignore suspicious patterns.</strong></p><p>Moscow’s precipitation records reminded me that unusual results deserve investigation. Sometimes, identifying a data-quality problem is more valuable than producing another chart.</p><p><strong>3. Break complicated problems into smaller ones.</strong></p><p>Whether you’re debugging Python code, writing SQL queries, or dealing with a difficult task outside programming, solving one part at a time can make the entire problem more manageable.</p><p><strong>4. Don’t confuse familiarity with understanding.</strong></p><p>Those four-minute quizzes made this very clear. Recognising a concept is not the same as knowing how to apply it. Practise independently, experiment, and test yourself without relying entirely on examples.</p><p><strong>5. Start before you feel completely ready.</strong></p><p>This was probably my most personal lesson.</p><p>With so much to learn and a strict deadline, it was tempting to wait until I understood everything before attempting the assessment.</p><p>But I realised that working through the tasks was itself part of the learning process.</p><p>Sometimes, you don’t become ready before you start. You become more capable by starting, making mistakes, and figuring things out along the way.</p><h3>Seven Days Later…</h3><p>At the beginning of the week, I genuinely wasn’t sure I could keep up with everything.</p><p>There was so much to learn and so little room to fall behind.</p><p>But somewhere between the NumPy arrays, SQL queries, database errors, weather analysis, and four-minute quizzes, I started seeing how the different concepts connected.</p><p>I also began to appreciate something beyond the technical skills.</p><p><strong>Learning data analytics isn’t just about knowing which tool or function to use. It’s about developing the ability to ask better questions, investigate problems, and make sense of information.</strong></p><p>If I could do one thing differently, I would start the assessment earlier, spread the learning activities more evenly across the week, and spend more time testing my understanding rather than simply reading examples.</p><p>And perhaps the most interesting part of the experience was how my relationship with SQL changed.</p><p>I started the week thinking it was my comfort zone.</p><p>I ended it realising there was still so much I didn’t know.</p><p>And surprisingly, that didn’t discourage me.</p><p>It made me love it even more.</p><p>For anyone else learning data analytics, I hope this serves as a reminder that struggling with a concept doesn’t necessarily mean you’re incapable of understanding it.</p><p>Sometimes, you’re simply encountering something new.</p><p>Keep experimenting. Ask questions. Check your results. Learn from the errors.</p><p>And don’t forget to enjoy those little moments when something that once looked confusing finally makes sense.</p><p>As for me, Week 6 was demanding, occasionally frustrating, and genuinely enjoyable.</p><p><strong>Seven days. Several new concepts. A few stubborn errors. Plenty of discoveries. And a much better appreciation of what it means to think like a data analyst.</strong></p><p>Now, if only the next set of quizzes could give us a little more time to think. 😂</p><p><em>Week 6 of my DataraFlow Data Science and Analytics journey.</em></p><p>#DataAnalytics #DataScience #SQL #Python #NumPy #Pandas #LearningInPublic #DataraFlow</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4caedf5c21d6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Thought I Needed a Research Topic. I Needed a Better Question.]]></title>
            <link>https://medium.com/@balqisbankole3/i-thought-i-needed-a-research-topic-i-needed-a-better-question-61916aa110d2?source=rss-52308e3b1ad9------2</link>
            <guid isPermaLink="false">https://medium.com/p/61916aa110d2</guid>
            <category><![CDATA[towards-data-science]]></category>
            <category><![CDATA[academic-research]]></category>
            <category><![CDATA[topics]]></category>
            <category><![CDATA[plant-science]]></category>
            <category><![CDATA[research]]></category>
            <dc:creator><![CDATA[Balqisbankole]]></dc:creator>
            <pubDate>Fri, 09 Oct 2026 11:07:54 GMT</pubDate>
            <atom:updated>2026-10-09T11:09:00.978Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*n9rHTbBOIBiD0c5OthGfrQ.png" /></figure><p>I thought finding a research topic would be easy.</p><p>Find an interesting problem. Find a dataset. Ask a question. Start working.</p><p>Simple, right?</p><p>Last week, I set out to do exactly that. I wanted to use publicly available data, build on my background in plant science, and find a new question to investigate without collecting data from the field.</p><p>Then things got complicated.</p><p>Sometimes, I found a dataset but could not think of an interesting question to ask. Other times, I had a question I liked but could not find data to answer it.</p><p>And whenever I found a promising idea, another question came up.</p><p>Has someone already done this?</p><p>Usually, someone had.</p><p>At some point, I started wondering if finding a research topic was a research project of its own. 😂</p><p>That experience taught me something important. A good research idea needs three things: a problem worth investigating, a question I can answer, and data that can help me answer it.</p><p>Finding all three at once is harder than it sounds.</p><p><strong>An idea sounds different when you have to explain it</strong></p><p>There is a certain comfort in having an idea in your head.</p><p>Everything makes sense. The connections are clear. You can almost see the finished research.</p><p>Then you open a document.</p><p>Suddenly, you have to explain the problem, why it matters, what other researchers have found, and what your own work will contribute.</p><p>I would start writing, realise something was missing, and go back to rethink the whole idea.</p><p>My brain: This is a brilliant research topic.</p><p>My document: Please explain.</p><p>That was when I watched a talk by Simon Peyton Jones called How to Write a Great Research Paper, presented by Microsoft Research.</p><p>I expected tips on how to structure a research paper. Instead, I got a different way of thinking about writing.</p><p><strong>What if writing helps you think?</strong></p><p>I used to think research followed a particular order.</p><p>First, choose a topic. Then, do the research and analyse the results. Finally, write everything up.</p><p>In other words, do all the thinking first and write about it later.</p><p>But the talk challenged that idea. Writing early can help you discover what you do not understand yet.</p><p>That sounded familiar.</p><p>While exploring crop-related datasets, I considered a question like this:</p><p>Which pepper accession has the highest nutritional value?</p><p>It seemed like a reasonable research question. Find the data, compare the accessions, and identify the best one.</p><p>Job done.</p><p>But then I started thinking about it more carefully.</p><p>What if one accession performs well for one trait but poorly for another? What if the best choice depends on whether a breeder prioritises yield, fruit quality, or nutritional content?</p><p>Simply identifying the accession with the highest value might not be enough to address a bigger problem.</p><p>I realised that I needed to think beyond what the data could show me. I also needed to understand why the answer would matter.</p><p>That is the difference between having an idea and having a question worth investigating.</p><p>Writing is not just the final step of research. It can also help you figure out what your research should be about in the first place.</p><p>And thankfully, I do not have to wait for all my thoughts to organise themselves perfectly before I start writing.</p><p><strong>A research paper is not a storage facility</strong></p><p>Another idea from the talk stayed with me: a research paper should have one central idea.</p><p>The goal is to communicate that idea clearly, not to include everything you have learnt along the way.</p><p>This is harder than it sounds.</p><p>You read one paper, which leads to another. That paper introduces a new concept, so you search for more information. Before you know it, you have opened fifteen tabs, saved twenty articles, and somehow forgotten the question you started with.</p><p>And because you spent hours reading, you want to include everything.</p><p>But a research paper is not a storage facility for everything you have learnt.</p><p>Knowing a lot about a subject and explaining your main point clearly are two different skills.</p><p>Readers should not have to wade through pages of background information before they understand what your research is actually about.</p><p>The same applies to a research proposal. If I cannot explain the problem and what I hope to contribute, adding more technical terms will not make the idea stronger.</p><p>Sometimes, the real work is deciding what to leave out.</p><p><strong>Before you analyse the data, ask yourself: why?</strong></p><p>When you find an interesting dataset, it is tempting to jump straight into the analysis.</p><p>You start imagining the charts, the patterns, and the models you could build.</p><p>But imagine spending weeks analysing thousands of observations and building a polished model, only for someone to ask:</p><p>So what? 😭</p><p>That would hurt.</p><p>A dataset is a resource, not a research question. And a complicated analysis does not automatically mean you have made a meaningful contribution.</p><p>I found a real example that helped me think about this differently.</p><p>Researchers studying Capsicum chinense, a type of pepper, published data on 120 accessions. The data included yield, plant and fruit traits, and capsaicin content.</p><p>One possible starting point is to ask which accessions perform best.</p><p>But what else could we investigate?</p><p>For example, do accessions with high yields also have high capsaicin content? Or are there trade-offs between traits that matter when choosing material for different breeding goals?</p><p>These are possible questions, not confirmed research gaps. I would still need to review existing studies and check whether the available data could answer them.</p><p>That is the point.</p><p>Finding a dataset is only the beginning. The real challenge is finding a useful question that the data can help answer.</p><p>I do not want to fall in love with a dataset before figuring out whether there is a meaningful story to tell.</p><p><strong>I am still looking for my answer</strong></p><p>Watching one video has not suddenly made me an expert in research writing.</p><p>I am still working on my proposal, questioning my ideas, and figuring out what the available data can realistically help me investigate.</p><p>But I am approaching the process differently.</p><p>When I find an interesting topic, I ask more than whether I like it. When I write, I pay attention to the questions that come up. And when I explain my research, I try to keep the main idea from getting buried under everything else I know.</p><p>Going back to rethink an idea does not always mean I am making no progress.</p><p>Sometimes, it means I am finally asking the right questions.</p><p>Even if it feels like staring at the same document for the third time.</p><p>Maybe the hardest part of research is not finding a topic. Maybe it is finding a question that deserves an answer.</p><p>If you have worked on a research project, which part tested your patience the most: finding a topic, making sense of the literature, or explaining an idea that made perfect sense in your head?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=61916aa110d2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How I Stopped Being Afraid of My Own Code]]></title>
            <link>https://medium.com/@balqisbankole3/how-i-stopped-being-afraid-of-my-own-code-fc6ec62c77e7?source=rss-52308e3b1ad9------2</link>
            <guid isPermaLink="false">https://medium.com/p/fc6ec62c77e7</guid>
            <dc:creator><![CDATA[Balqisbankole]]></dc:creator>
            <pubDate>Sat, 26 Sep 2026 12:23:45 GMT</pubDate>
            <atom:updated>2026-09-26T12:23:45.554Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>The Error Message That Used to Scare Me</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*19cc2FZaLVcKRbwDsQliLg.png" /></figure><p>There is a particular moment I have started to recognise when learning Python.</p><p>I write a few lines of code. I press run. For a few seconds, everything looks fine. Then the screen changes.</p><p>A traceback appears. A line number. An error type. Some words I recognise, some I don’t.</p><p>For a long time, that was enough to make me feel like I had done something terribly wrong. I would read the message from the top, then the bottom, then somewhere in the middle, hoping the answer would reveal itself. It rarely did. Instead, I would start wondering whether I had missed something obvious, whether this was one of those parts of programming that would always feel foreign to me.</p><p>Then something changed. I started becoming curious about the errors. Not <em>why am I getting this wrong</em>, but: <em>what is Python actually telling me?</em></p><p>That was a much better question.</p><h3>When the Screen Turned Red</h3><p>Python errors are not random messages. A traceback shows where an exception occurred, and the final line identifies the type of exception and usually explains what happened. I had been looking at all that information as intimidation. I was beginning to see it as information.</p><h3>You Don’t Know What You Don’t Know</h3><p>I hit this with dates. I had a date stored as text and needed it as a proper datetime value. I didn’t know the right way to do it, so I guessed. I tried wrapping the string directly in datetime(), something like datetime(&quot;2024-03-15&quot;), assuming Python would just figure out what I meant.</p><p>It didn’t. I got a TypeError, and for a second the old instinct kicked in: the screen turns red, and some part of me still wants to read that as <em>you&#39;ve done something wrong</em>.</p><p>But this time I asked a different question. Not <em>why is this broken</em>, but <em>what does Python actually offer for this?</em></p><p>So I searched. I remembered something I’d recently discovered: dir() lets you explore the names available in an object or module. Instead of <em>I don&#39;t know this, so I&#39;m stuck</em>, I could ask: <em>what is available to me?</em></p><p>That led me to strptime(). datetime.strptime() parses a date string according to a specified format and returns a datetime object. Suddenly the problem had a name, and the error made sense too: I hadn&#39;t written broken code, I&#39;d just been using the wrong tool for the job.</p><p>You are not always stuck because the problem is too difficult. Sometimes the error itself is just telling you that you haven’t found the right tool yet, and that’s a much less frightening thing to be told.</p><h3>Teaching the Code What to Do When Things Go Wrong</h3><p>That same idea came back when I started working with exceptions. I had always thought of an error as something that happened to me: I wrote the code, Python rejected it, I panicked.</p><p>But exception handling offered a different way of looking at it. What if the program doesn’t have to stop every time something unexpected happens? What if I can decide how certain problems get handled?</p><p>That’s where try/except started making sense. Python runs the code inside try; if an exception occurs, it looks for a matching except clause:</p><pre>try:<br>    number = int(input(&quot;Enter a number: &quot;))<br>    result = 100 / number<br>except ValueError:<br>    print(&quot;Please enter a valid number.&quot;)<br>except ZeroDivisionError:<br>    print(&quot;You cannot divide by zero.&quot;)</pre><p>Now I can handle two different situations separately instead of catching everything with one broad except. I could tell the program: if this happens, respond this way; if that happens, respond another way. Handling errors isn&#39;t about pretending problems will never happen. It&#39;s about deciding how the program responds when they do.</p><p>Then came else, which runs when the try block completes without raising an exception:</p><pre>try:<br>    number = int(input(&quot;Enter a number: &quot;))<br>    result = 100 / number<br>except ValueError:<br>    print(&quot;Please enter a valid number.&quot;)<br>except ZeroDivisionError:<br>    print(&quot;You cannot divide by zero.&quot;)<br>else:<br>    print(&quot;The answer is:&quot;, result)</pre><p>Something goes wrong, I have a response. Nothing goes wrong, I have a response for that too.</p><p>Then there was raise, which lets you deliberately trigger an exception:</p><pre>age = int(input(&quot;Enter your age: &quot;))</pre><pre>if age &lt; 0:<br>    raise ValueError(&quot;Age cannot be negative.&quot;)</pre><p>Python can represent -5 perfectly well as an integer, but if my program&#39;s rules say age must be non-negative, I can decide that value is invalid and raise on it myself, rather than letting bad data quietly flow downstream and fail somewhere less obvious later.</p><p>I was starting to see that good code isn’t only written for the version of events where everything goes right. It’s written with some imagination for what could go wrong.</p><p>That kind of thinking turned out to matter earlier than I expected, before I’d even written a line.</p><h3>Learning to Pause Before I Start</h3><p>Sometimes the hardest part isn’t writing the code, it’s knowing where to begin. I’ve looked at a task, understood the English sentence perfectly, and had no idea what the first line of Python should be. My instinct used to be to start typing anyway. Sometimes that worked. Other times I created a problem before I’d even understood the problem.</p><p>I’m learning to pause first: What am I being asked to do? What information do I have? What do I need to produce? What could go wrong? What tools might help? Then I start.</p><p>Even the terminal has lost some of its mystery. Creating a virtual environment, activating it, installing a package, checking what’s installed, none of it feels as intimidating once you know where you are, where you want to go, and what command gets you there.</p><p>None of this makes the errors disappear. It just means fewer of them catch me off guard, and the ones that still do, I know what to do with.</p><h3>The Error Message Is Still There</h3><p>The error messages are still there. The traceback is still there. My code still breaks. But I don’t look at those messages the same way anymore.</p><p>Maybe progress isn’t becoming someone who never gets errors. Maybe it’s becoming someone who knows what to do when they appear.</p><p><strong>#Python #Programming #LearningToCode #SoftwareDevelopment #BeginnerCoding #SoftwareDevelopment #DataScience #TechWriting #PersonalGrowth #100DaysOfCode #WomenWhoCode #CareerChange #SelfTaughtProgrammer #PythonProgramming #LearnToCode</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fc6ec62c77e7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Week Python Started Making Me Ask “But Why?”]]></title>
            <link>https://medium.com/@balqisbankole3/the-week-python-started-making-me-ask-but-why-01e6ae6bdff6?source=rss-52308e3b1ad9------2</link>
            <guid isPermaLink="false">https://medium.com/p/01e6ae6bdff6</guid>
            <dc:creator><![CDATA[Balqisbankole]]></dc:creator>
            <pubDate>Thu, 17 Sep 2026 16:18:38 GMT</pubDate>
            <atom:updated>2026-09-17T16:18:38.959Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Fde8E9H3nxY_gx8m" /><figcaption>Photo by <a href="https://unsplash.com/@jstrippa?utm_source=medium&amp;utm_medium=referral">James Harrison</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>I thought I was fairly comfortable with Python before this week.</p><p>I have used it for data analysis and machine learning, so I was familiar with writing Python code and getting things done with it. Then this week at DataraFlow, we went deeper into Object-Oriented Programming and some parts of Python that I had not really had to think about before. Some concepts were straightforward, but others made me stop and ask questions that I had never really considered before.</p><p>And I think those were the most interesting parts of the week.</p><p><strong>The __str__() Question I Didn’t Know I Had</strong></p><p>One small thing that caught me was the difference between __init__() and __str__(). __init__() started to feel very familiar because almost every class example used it to initialise the object’s attributes. But then I came across __str__() and my first thought was basically, “Okay, but when am I actually supposed to use this?”</p><p>I understood what it did technically. Without __str__(), printing an object gives you Python’s default representation, while defining __str__() lets you control what the object looks like when it is converted to a string. What I found less obvious was knowing when that was actually useful. Why would I need to customise the way an object is displayed instead of just accessing its attributes directly?</p><p>The difference became clearer when I looked at the purpose of each method. __init__() is about setting up an object when it is created, while __str__() is about giving that object a useful human-readable representation when it is displayed. So if I create a Person object with a name and age, __init__() can store those values, while __str__() can decide that printing the object should give me something like Balqis(25) instead of Python’s default object representation.</p><p>That distinction seems small, but it helped me understand that these special methods are not just things Python expects me to put inside every class. They each have a specific purpose. __init__() is commonly used when an object needs to be initialised with data, while __str__() becomes useful when I want the object itself to have a meaningful string representation.</p><p>That was one of those moments where I realised I didn’t just need to know what a method does. I also needed to understand when it is useful.</p><p><strong>Then Iterators Happened</strong></p><p>Iterators were probably where things became properly challenging.</p><p>I have written plenty of for loops without thinking too much about what happens underneath. Give Python a list, loop through it, and move on. But once I had to understand the iterator protocol and work with __iter__(), __next__() and StopIteration, the simple for loop suddenly had a lot more going on behind it.</p><p>The Fibonacci iterator task made this even more interesting. Generating Fibonacci numbers isn’t particularly complicated, but creating a class that produces one number at a time requires the object to keep track of its current state. It also needs to know when it has reached the end and raise StopIteration to signal that there are no more values.</p><p>The part that helped me was understanding the difference between an iterable and an iterator. An iterable is something Python can get an iterator from, while an iterator is responsible for producing the next value. Once there are no more values, StopIteration tells Python that the iteration is finished.</p><p>It sounds much simpler when you explain it afterwards.</p><p>While learning it, though, I had to go over the concept more than once. When the course explanation didn’t completely click, I looked for another explanation on YouTube and then came back to the topic. Seeing the same idea explained differently helped me connect what I was reading to what the code was actually doing.</p><p>Now when I see a for loop, I know there is more happening underneath it than what I can see on the screen.</p><p><strong>Polymorphism Made More Sense When I Saw the Point</strong></p><p>Polymorphism was another concept that initially sounded more complicated than it actually was. The definition is easy enough to read, but I found it more useful to think about why I would want it in the first place.</p><p>One of the tasks involved creating Rectangle, Circle and Triangle classes, each with an area() method. Each shape calculates its area differently, but I can put the different objects into one list and call area() without checking which type of object I am dealing with.</p><p>That was the part that made the concept click.</p><p>The code doesn’t need to ask, “Are you a rectangle? Are you a circle? Are you a triangle?” It simply calls area(), and each object knows how to respond. The same method name can therefore produce different behaviour depending on the object.</p><p>That made polymorphism feel much more practical than simply memorising the definition. It showed me how OOP can allow different objects to work with the same piece of code without needing a separate instruction for every possible object type.</p><p><strong>Then Python Made Me Think About MRO</strong></p><p>Multiple inheritance was another topic that required more attention.</p><p>If a class inherits from two parent classes and both parents have a method with the same name, which method does Python use? That question led to Method Resolution Order, or MRO, and this was one of those concepts where I didn’t want to simply run the code and see what happened. I wanted to understand why Python chose one method over another.</p><p>The Duck example with Flyer and Swimmer made the idea easier to see. If both parent classes have a move() method, Python follows a specific order when searching for the method. Once I understood that order, the behaviour stopped looking random.</p><p>I liked that part because it changed how I approached the code. Instead of waiting for Python to tell me what happened, I could look at the class structure and reason about what would happen before running it.</p><p><strong>The More I Learnt, the More the Concepts Connected</strong></p><p>What made this week interesting was that the concepts didn’t stay completely separate.</p><p>At first, __iter__(), __next__(), polymorphism, inheritance and MRO were just different things on a list of topics. But as the tasks became more involved, I started seeing why each concept existed and how they could work together.</p><p>That was probably the biggest difference between this week and simply reading about Python concepts online. It is easy to read a definition and think, “Yes, I understand that.” It is another thing to see a piece of code and understand why Python behaves the way it does.</p><p>And when I didn’t understand something immediately, I had to find another way into it. Sometimes that meant watching another explanation. Sometimes it meant going back to a simpler example before returning to the actual task. Sometimes it was just giving myself enough time to process what I had read.</p><p>I definitely had moments this week when my brain was tired. 😂 But I also had a few of those satisfying moments where something that had looked unnecessarily complicated suddenly made sense.</p><p><strong>What I’m Taking Away From This Week</strong></p><p>I don’t think the biggest lesson from this week is that Python is difficult. It is that there is a lot more depth to a language you may already feel comfortable using.</p><p>I could use Python before this week. I could write code and work with data. But now I’m starting to pay more attention to what Python is doing underneath that code and why certain features exist.</p><p>That has changed the kind of questions I ask.</p><p>Instead of only asking, “What does this do?”, I’m starting to ask, “Why does Python work this way, and when would I actually use it?”</p><p>And I think that is where learning starts getting more interesting.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=01e6ae6bdff6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My First Week at DataraFlow: What I Learnt About Data Science and Getting Stuck]]></title>
            <link>https://medium.com/@balqisbankole3/my-first-week-at-dataraflow-what-i-learnt-about-data-science-and-getting-stuck-29a014edf39a?source=rss-52308e3b1ad9------2</link>
            <guid isPermaLink="false">https://medium.com/p/29a014edf39a</guid>
            <category><![CDATA[python-programming]]></category>
            <category><![CDATA[vscode-theme]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[learning-to-code]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Balqisbankole]]></dc:creator>
            <pubDate>Wed, 02 Sep 2026 13:53:56 GMT</pubDate>
            <atom:updated>2026-09-02T13:53:56.435Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>Week One at DataraFlow: Learning to Figure Things Out</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eUx11ErOXrKOQH5o5nMfxQ.png" /></figure><p><strong>Getting in is the easy part.</strong></p><p>That was one of the first things I heard during my orientation at DataraFlow, and it turned out to be a good way to describe my first week.</p><p>I started the week with orientation, introductory sessions on data science, reading materials, software installations, a crash course on Python, and a series of tasks and assessments.</p><p>It was a busy first week, but more than anything, it made me think differently about what it actually means to learn.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mJPBo81ipCkyOwtXYCgqTw.png" /></figure><p><strong>So, what is data science?</strong></p><p>One of the first things I did this week was go back to the basics and ask myself what data science actually means.</p><p>I already knew that it involved data, statistics, programming and different analytical tools. But the introductory sessions helped me think about it less as a collection of technical skills and more as a way of solving problems.</p><p>Businesses generate data constantly. The important part is not simply having that data, but being able to use it to understand what is happening, find patterns, answer questions and support better decisions.</p><p>That can involve collecting and cleaning data, exploring it, analysing it, visualising it and communicating what you find.</p><p>This made something from the orientation particularly relevant:</p><p>The problem you solve is the product. The tools are not the product.</p><p>It is easy to become focused on learning Python, SQL, R or any other tool. But the real goal is being able to use those tools to solve something meaningful.</p><p>That is one of the things I hope to develop throughout this internship: learning how to approach real business problems, not just learning how to use software.</p><p><strong>The orientation changed how I think about learning</strong></p><p>The orientation was probably the part of the week that stayed with me the most.</p><p>We talked about consistency, ownership and what happens when you do not know what to do.</p><p>One idea that stood out was that potential expires.</p><p>Having the ability to do something is not enough if you never develop it. And being the smartest person in the room does not necessarily mean you will be the one who succeeds.</p><p>Sometimes, consistency matters more.</p><p>We were also introduced to the idea of becoming a “figure it out” person.</p><p>This made sense to me because the tools we use in technology are constantly changing. What matters is not knowing every tool forever, but being able to learn a new one when you need it.</p><p>The stack will change.</p><p>The ability to learn should not.</p><p><strong>My first challenge: getting things to work</strong></p><p>One of my first challenges was setting up the tools required for the programme.</p><p>Some of the installation involved using the terminal, and I had to get more comfortable with entering commands, understanding what the terminal was telling me and figuring out why something was not working.</p><p>It was frustrating at times, especially when I just wanted to move on to the actual learning.</p><p>But it gave me an opportunity to practise something we had discussed during orientation: read the error message properly.</p><p>Instead of immediately seeing an error as a dead end, I started treating it as information.</p><p>What exactly went wrong?</p><p>What was the message telling me?</p><p>What could I try next?</p><p>That small change in approach made the problems easier to deal with.</p><p><strong>Then came Python</strong></p><p>The Python crash course was another major part of the week.</p><p>I went through the foundations, including variables, data types, operators, input, conditions, loops, lists, tuples, dictionaries, functions, modules, exceptions and file handling.</p><p>The concepts made sense while I was learning them, but writing actual programmes was where things became interesting.</p><p>Sometimes I knew exactly what I wanted a programme to do but struggled to translate that idea into code.</p><p>Other times, I could get the code to run but realised that it was not behaving exactly the way I wanted.</p><p>For one of the programmes I worked on, I wanted the user to enter a number within a particular range. After getting that part to work, I realised there was another problem with the way I had written the script.</p><p>The programme allowed the user only five attempts, but that included attempts where the user entered something outside the required range.</p><p>What I actually wanted was for the user to have five valid attempts, regardless of how many invalid values they entered along the way.</p><p>So I had to go back to the logic and change the way the programme counted the attempts.</p><p>It was a small difference in what I wanted the programme to do, but it made me realise how important it is to think through the behaviour of a programme, not just get the code to run.</p><p>Getting the code to work is one thing. Getting it to work the way you intended is another.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SgTei-Jj75EMeonCMHTjVA.png" /></figure><p><strong>Learning how to be stuck</strong></p><p>I think this was probably my biggest takeaway from the week.</p><p>The orientation gave us a simple process for dealing with problems:</p><p>Break it down. <br>Search properly. <br>Try something. <br>Ask. <br>Document it. <br>Try again.</p><p>I like this because it turns being stuck into something I can actually respond to.</p><p>There was another idea that I found particularly useful:</p><p>“I don’t get it yet” is a timeline. “I don’t get it” is an identity.</p><p>There were definitely moments this week when I did not understand something immediately.</p><p>But that does not mean I cannot understand it.</p><p>It means I have not understood it yet.</p><p><strong>What I want from this internship</strong></p><p>I am not joining this programme just to add more tools to my CV.</p><p>I want to become better at solving problems with data.</p><p>I want to work with real business problems and understand how data can be used to answer questions, identify issues and support decisions.</p><p>I also want to get better at showing my work.</p><p>Instead of simply saying “I learnt Python this week,” I want to be able to show what I built, what went wrong, what I tried, what I learnt and what I would improve.</p><p>That is part of why I am writing these weekly reflections.</p><p><strong>One week down</strong></p><p>My first week at DataraFlow has been a combination of learning, experimenting, getting stuck and trying again.</p><p>I have learnt more about data science, started strengthening my Python foundation, worked through practical tasks, set up new tools and completed my first assignments and assessments.</p><p>But beyond the technical skills, I think I am beginning to understand that learning technology is not about always knowing the answer.</p><p>It is about knowing what to do when you don’t.</p><p>The tools will change.</p><p>The problems will change.</p><p>The technology will change.</p><p>So I need to keep learning.</p><p>For now, I am taking things one week at a time, one problem at a time and, hopefully, becoming a little better at figuring things out each time I get stuck.</p><p>One week down. Many more to go.</p><p>What is one thing you learnt when you were starting something completely new?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=29a014edf39a" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>