JBarcelona 2023
The challenge
This challenge was the problem of a Nuwe competition. The goal of the challenge was to build a classification model which had to predict which costumers decide to end their contracts with the company we get the data from, so the percentage of people who stop being clients decreses by analyzing the reasons.
The approach
Since this was my first time building a model, I decided to start by looking at some guides and understand what had to exactly do the model. I manage to get something and predict some results, and it was really nice when I checked and turn out I had a f1-score of 0.9. Then I also tried some optimitzations which didn’t work, but that was enough to pass to the final.
Metrics of each trained model.
Project’s code on Github
Google Colab
The final
When I arrived at the place the final was, I met my team for the first time (as the previous round was individual) and we were people from differents backgrounds. All of us collaborated in order to solve the final challenge in which we created a website using streamlit that could be use to visualize women’s football statistics and analysis.
Team view on the website.