Q&A with Denys Bespalko: Creating a data-driven recruitment tool for NAC Breda

Q&A with Denys Bespalko: Creating a data-driven recruitment tool for NAC Breda

05/23/2024 - 12:51

We sat down with Denys Bespalko, a first-year student in the Applied Data Science & Artificial Intelligence (ADS&AI) study programme at Breda University of Applied Sciences (BUas), to discuss his project for the football team NAC Breda.

Denys developed a data-driven tool to help the club make strategic decisions about hiring new players. Here’s what he had to say about his project and his experience.
Data Science & AI
  • Stories
  • Student work

Can you describe the project that you have been working on? 

Denys: ‘I’ve been working on a recruitment tool for NAC Breda, designed to help them find new players and manage emergency transfers. The tool uses data-driven design to customise player selection based on the club's needs and budget. It considers parameters like age, position, and other performance metrics. The idea is to guide the club in making strategic decisions about player acquisitions.’  

‘We had a chance to visit NAC Breda, where we were guided by a member of their data handling team. He provided us with the necessary data and explained their requirements. Throughout the project, we had a mid-term meeting to present our findings and receive feedback, which helped me refine my approach.’ 

What did the process of this project look like? 

Denys: ‘The process started with NAC introducing us to their data management practices, including how they handle personal information about players and their recruitment data from other European teams. We had to choose between using regression or classification models. Regression could help predict continuous variables like a player's potential market value, while classification could categorise players into specific groups based on their value.’  

‘I chose a different strategy by treating each player as a vector in a multi-dimensional space with features like goals scored, successful defenses, accurate passes, and tackles. Using cosine similarity, I compared each player's vector to find the closest match based on various criteria. This method allowed the tool to suggest the best player matches for NAC Breda.’  

‘After processing and visualising the data, we analysed the correlation between different features and the performance metrics. The outcome was a report that included the problem statement, business value, model details, ethics considerations, and recommendations for NAC Breda.’ 

What are you most proud of? 

Denys: ‘I’m most proud of finding my own unique approach to the project. It was interesting to see how my peers tackled the same problem, especially since I didn’t know them well at the start of the year. My project turned out to be the best, which was very rewarding.’ 

‘I also created a simple website for the tool’s user interface. Although it’s basic, I’d like to improve the UI and perhaps develop an API to make it more visually appealing and user-friendly.’  

What are your future plans? 

Denys: ‘This project provided a great foundation for developing my skills further, and I’m excited to build on this experience.’ 

  

To learn more about Denys Bespalko and his future projects, you can connect with him on LinkedIn