(written by Hanne Van Schepdael- element61 Trainee in 2024)
Who are you and what are your educational goals?
My name is Hanne Van Schepdael, and I am currently pursuing a master’s degree in Business Engineering with a specialization in Data Analytics at Ghent University. To gain practical experience and apply my theoretical knowledge to real-world business cases, I applied for a summer internship in 2024. My primary goal was to see how the frameworks I studied in university are applied in a professional context.
What did your internship look like? What did you do exactly?
The central question guiding my 6-week internship was to uncover root causes in specific scenarios, such as identifying anomalies, peaks, or downturns. For instance, “How can fluctuations in revenue and product sales be analyzed to pinpoint the root causes of discrepancies in demand forecasting models?” A key sub-question I aimed to answer was how to estimate causal relationships among various variables. By estimating these cause-effect relationships within the model, I could identify which variables most significantly impacted the outcome.
I spent most of the time of my internship on developing an automated method to calculate cause-effect estimates between variables, with as an input a demand forecasting basetable. The output was intended to be an estimate per pair of variables. To achieve this, I researched and experienced with dozens of Python packages. In the end, it turned out that most of these Python packages were too ‘advanced’ in the sense that it wasn’t provided to work on all sorts of datasets. I also encountered several other limitations such that I concluded the following. Causal inference cannot be fully automated with these methods and always requires domain expertise to understand the causal relationships among variables. Additionally, I found that while estimating causal inference could aid in root cause analysis, it is not sufficient on its own. To fully understand the root causes of scenarios such as anomalies, peaks, or declines, additional information is necessary.
I also proposed an alternative approach to my primary research question by creating a visual analysis of sales trends in Power BI. For the case I studied, I broke down sales by product and store, filtered by time, providing a clearer view of sales trends and aiding in identifying root causes behind peaks or declines.
In addition to my project, I engaged with the broader work at element61 by attending meetings and talking with consultants. It was fascinating to learn about the types of business cases they tackle and their methodologies.
How did you experience the contact with the team of element61?
During my internship at element61 I felt very welcome by the team. Everyone was open to try the best they could to answer my questions. I also got encouraged to ask all the questions that came to my mind and involving as many people as possible (from different teams) to optimize my learning experience. For them, it was very important that I didn’t waste my time and that I learned a lot during my time there.
What is your biggest takeaway from the internship?
This internship was invaluable both personally and professionally. I gained confidence and developed skills that will benefit my future career. I also learned that even if research outcomes differ from expectations, the process can still contribute valuable insights to the bigger picture. On a technical level, I have been able to improve my Python skills and was introduced to tools like PySpark and Microsoft Power BI, which will be valuable as I complete my degree.
What do you want to share with any future students considering an internship at element61?
For hands-on experience in data analysis and engineering, element61 is an excellent choice. The guidance you receive helps you develop autonomy and learn to ask the right questions to move forward independently. I’m also grateful for the chance to make a meaningful contribution through my research topic and explore practical applications.
In short: a fantastic opportunity!