From Print(“Hello”) to Data Analysis: My Thesis with AI-Assisted Coding
9
October
2025
No ratings yet.
When I started my thesis, I barely remembered how Python worked. I knew what a dataset was and how to print a line or write a simple loop, but that was about it. The idea of building an entire data-science workflow seemed far beyond what I could do on my own. Yet, a few months later, I had written a full pipeline to analyze hybrid work patterns using behavioral logs, location data, and daily surveys. What made that possible was Generative AI.
ChatGPT quickly became my silent collaborator. Whenever I got stuck, I simply described what I needed: filtering AWT data by time, merging JSON files by date, or running a Mann-Whitney U-test. Within seconds, it generated structured and readable code that actually worked. It helped me clean and merge datasets, calculate metrics like active work time and task switches, and even combine GPS data with behavioral data to label each day as home or office. Suddenly, something that felt completely out of reach became manageable.
Of course, the process was not perfect. I often had to debug the AI’s mistakes, rewrite lines of code, and verify that the logic fit my data. Sometimes ChatGPT used outdated Pandas functions or made assumptions that didn’t make sense. But those moments taught me more than any tutorial could. I started to understand not just what the code was doing but why it worked that way.
Looking back, Generative AI didn’t write my thesis for me; it expanded what I was capable of. It turned Python from something intimidating into a tool I could actually use. For me, that is the real power of AI. It doesn’t make you less of a coder; it makes you more confident to learn, experiment, and create things you once thought were impossible.
The Unfulfilled Promise of an AI that can take my Job
30
September
2025
5/5 (1)
With a background in Computer Science, I was able to enter the job market as a software engineer early on. This way I started to work as a programmer at a medium-sized Dutch software company after my first year of studying as a Bachelor. At that time, AI and Generative AI would not have an impact on our line of work for another one and a half years when ChatGPT 3 would launch for the first time.
When working on a large enterprise system for industries with unique and complex processes the complexity in software architecture and class structure increases exponentially. Where in class a coding exercise might have entailed creating a few classes, implementing a few constructors and running a specified set of methods all within a predefined programming language, coding at a software company involves countless additional steps. Even the simplest bug fixes and feature developments require deep knowledge over how a niche subset of the source code functions, a great ability in reading and understanding complex calculations/algorithms ran by the backend and intricate knowledge in not just multiple coding languages (Java, JavaScript, TypeScript…), but also frameworks that run on these languages such as React, Node or Ember.
It was no surprise therefore that all of us were quite intrigued by the potential of AI-assisted coding right from the release of ChatGPT 3. Coding plugins and extensions built into the Integrated Development Environments (IDEs) were already widely utilized within the company for many years and they helped focus on the underlying logical fallacies that need to be solved. With the new Generative AI, however, the premise was that the assistant could assist or take over even this work. After much experimentation and the implementation of an enterprise version of Google Gemini we quickly reached the limits of AI’s coding capabilities in today’s world. After so much drama in the news and a public perception of AI as the coding-killer we found that although Gemini could analyze and correct a couple of individual lines of code, it is not yet able to navigate or store a large codebase to analyze and understand the context of the problem or feature.
Even companies like Microsoft, Google and Meta, who are some of the only organizations on earth able to afford to train their own Gen AI models on their own code, are unable to rely on their AI to fix small bugs autonomously. Too much risk is involved in incorrect design choices, edge case bugs and most importantly the verification process. This process is critical and still requires testing by real humans, who are skilled and competent enough to assess the end results based on chosen requirements.
For us and the rest of the development world, AI coding assistants will stay limited to “chunking” code into deliberately chosen fragments, selected by the developer, that can help the AI in assessing a coding task. A great improvement that can yield automatic generation of “boilerplate code” (repeated code that is commonly used throughout a project), the generation of a “stub implementation” to go off or a list of suggested corrections in case a developer gets hard stuck. Though, generative AI does not nearly constitute a true job killer and even if it could, it will take additional years until full capabilities will be available to the large majority of software companies on earth.
Coding and GenAI – an ideal match
8
October
2024
No ratings yet.
I remember my first assignment with RStudio in the Business Information course in the first year of my bachelor’s (in 2021). I had no experience with coding so I had to learn this from scratch. After crying and asking my lovely dad for help, I managed to write some code. This was all fun and games until I encountered my very first error code – and soon after many more followed. So of course, I stressed out and I didn’t know what to do. I started searching on Google and clicked on a number of webpages. And before I knew it, I was in deep into Stack Overflow. When I finally managed to find a post describing the same problem I encountered, I figured out that that the specific solution did not entirely work for my code, so the cycle repeated itself again. This process of resolving errors was time-consuming and definitely frustrating going from webpage to webpage and then going back to my code to see that the solution did not apply to my situation.
My journey
In 2022, ChatGPT entered the scene. Back then, I was on exchange in Hong Kong. While calling my dad, he mentioned something about an “AI bot who can generate everything you want it to generate”. I brushed this off seeing it as something similar to the Metaverse: the hype will last for a couple of months and then it will just be over. But when I came back home months later, everyone was still talking about this ‘ChatGPT’. One day, I tried putting my error codes in ChatGPT and it came up with a clear solution and even provided example code! How convenient! Now, I still use ChatGPT to take a look at my error codes.
Reflection
As you, the reader, might have noticed, I am very satisfied with using GenAI for coding. As I now genuinely love coding in RStudio, I still write my code myself. However, when encountering error codes, ChatGPT is definitely my ‘bestie’. Solving an error code is not the scope of writing code or running analyses. By using GenAI, it saves me a ton of time. No need to dive into the deep rabbit holes of Stack Overflow. I just provide my code and the error code to the GenAI bot and it will help me instantly. How amazing! It still blows my mind.
You still need to use your brain!
Using GenAI for coding does not, however, mean that you just let ChatGPT write your code and you just shut off your brain. In my opinion, it is still crucial to use a critical attitude when using GenAI for coding purposes. You can let the bot write all the code you want and resolve all of your problems every time, but if you don’t understand what the output is, using GenAI might as well be pointless. In my opinion, learning from the output from GenAI is the key to success. By learning from the output, you can prevent the same error codes next time you write code.
Thus, in my opinion GenAI is definitely useful for coding. No more wasting time on Stack Overflow as error codes are resolved in just a couple of seconds thanks to GenAI. Combine this with maintaining a critical mind and you might be able to hack a government soon.
Questions for the readers
As readers, are you as enthusiastic as me about using GenAI for coding? What are your experiences? Or have you encountered any difficulties by using GenAI for coding?
Generative AI in the Data Scientist’s Universe: A Friendly Companion or a Sneaky Enemy?
2
October
2023
5/5 (1)
In today’s business landscape, data is key. The widespread adoption of Artificial Intelligence (AI) and Machine Learning (ML) has sparked a compelling debate about the future of the data scientist’s role. As we progress into an era where being code literate and possessing technical knowledge are no longer barriers to entry, the accessibility of generative AI tools has transcended the boundaries of academic research and tech giants. It’s now within reach for anyone with an internet connection and an email address.
It made me wonder if this transformation means the death of data scientist-like professions or whether it signifies the birth of new roles. I will delve into this question using my personal experience as a starting point.
The Learning Curve
About a year ago I started in a newly created role as a junior business controller. I quickly realized the importance of harnessing data to create insightful dashboards that guide data-driven decision-making. Learning to navigate the complex world of data was a challenge in itself (I mean, how do I know that I can trust my data? What data is available and what is relevant?), but the real game-changer came when I discovered the power of generative AI.
It was like having a 24/7 coding mentor at my fingertips
Starting from scratch with SQL was no walk in the park. I encountered countless errors, syntax hiccups, and moments of sheer frustration. But I followed some courses, watched YouTube videos and found out that SQL is amazing (honestly: I would highly recommend learning it)! It was during this process that I stumbled upon an invaluable ally: ChatGPT. Whenever I hit a roadblock, I would turn to ChatGPT and put my code-related questions in. Within seconds, I’d receive clear and concise explanations, troubleshooting tips, and even code snippets to resolve my issues. It was like having a 24/7 coding mentor at my fingertips.
ChatGPT can even help write queries from scratch! Source: https://blog.devart.com/how-to-use-chatgpt-to-write-sql-join-queries.html
The Role of Generative AI in Data-Related Work
Generative AI, like ChatGPT, is not just a tool for beginners like me. I’m convinced that it has the potential to revolutionize the way data scientists and coders work. But should we see it as a friendly companion or a sneaky enemy that is stealing jobs?
AI as a Productivity Booster
First and foremost, generative AI can significantly enhance the productivity of data professionals. It excels at automating repetitive tasks, such as data retrieval and cleaning as well as tackling error messages. This frees up valuable time for more critical analysis. It can also serve as a valuable learning resource, providing instant answers and explanations for coding queries. Due to this I have experienced a steep learning curve in my SQL journey.
Human data scientists and coders bring a unique skill set to the table; they possess the ability to interpret, contextualize, and make nuanced decisions based on data
The Future of Data-Related Jobs
The future of data-related jobs stands at a crossroads, where the demand for AI-ready professionals who can effectively harness tools like ChatGPT is already evident in job descriptions. The fear of AI replacing human expertise is a legitimate concern, but it’s important to remember that AI is a tool, not a replacement. While generative AI offers substantial productivity gains, it’s not immune to errors or hallucinations, reminding us of the irreplaceable value of deep domain expertise. Human data scientists and coders bring a unique skill set to the table; they possess the ability to interpret, contextualize, and make nuanced decisions based on data. AI can undoubtedly assist in the technical aspects of the job, streamlining processes and automating tasks, but it cannot replicate the human touch required for holistic and insightful data analysis.
Well, the way I see it..
In my role, generative AI has truly been a game-changer. It’s become an ally helping me out with coding challenges and making my dashboards way better and more trustworthy. It surprised me by getting results faster than I expected, and I can’t complain about that! And whenever my SQL query was actually running properly, I put it in and asked for suggestions to improve performance. Using ChatGPT for my coding-related questions has boosted my efficiency and effectiveness, and I see it as a sneak peek into the future of data work. It’s all about humans and AI teaming up, playing to each other’s strengths for some seriously awesome outcomes.
So, the next time you encounter a coding conundrum, don’t hesitate to turn to generative AI. It’s not a replacement for your expertise but a trusted partner in your data-driven journey.
Curious to hear your thoughts and experiences! Also let me know whether some of you use other GAI tools for your data- and coding-related issues! 😉
Should we start teaching coding since the first years of education?
6
October
2021
No ratings yet.
We live in a society where data and technology are everywhere and are becoming an essential part of our lives. Companies are using big data in order to acquire competitive advantage and improve their process and products. Some observed benefit that data driven companies are enjoying compared to those who are not are:
A likelihood 23 times higher to acquire new customers thanks data driven marketing campaign
a profit increase of 8%
reduction of cost of 10% (keboola, 2019)
The demands from employer of coding and data analytics skills is skyrocketing. Glassdoor reported that eight of the top 25 jobs in the US are tech-based and require some level of coding proficiency. Similarly, a 2016 Burning Glass report found that the demand for roles such as data analyst is rising 12% than the market average (Nord Anglia 2020).
Reading this data one question came immediately to my mind. Since schools have a responsibility to provide students with all the tools and skills they need in order to succeed in their future should the educational system be more prone to the digital revolution and start teaching codes and data analytics since the first years of instruction? I personally believe that our educational system should empathize more the importance of those skills since the earlier age of our education.
When I started my master in Lisbon I had one mandatory course that was about econometrics and data analytics. So, the first time I started learning coding was at 22 years old and I wish I could have started way before. During my bachelor in management in Padova I did not have the possibility to choose any elective course regarding those topics. Many italian business university still do not have a dedicated bachelor or master to business and data analyst. I believe every students in management , finance or economics nowadays should have at least a basic knowledge in coding and data analytics in order to be competitive in the job market in the future. Like with every other subject if we start to approach it since we are young it would be easier to learn it and master it.
Let me know what you think and how is the situation in your countries.
