Read more about the project here: https://danilogb.github.io/2023/06/04/mindprober.html
This repository contains the code and documentation for the MindProber Sports Analysis project, which was completed as part of a soft skills and career development program at the University of Porto in March 2023.
The aim of the project was to investigate the correlation between MindProber's proprietary emotional engagement metrics and sports events, specifically focusing on the 2022 FIFA World Cup. MindProber is a local Portuguese startup that combines remote sensing and machine learning to deliver objective and comparable emotional engagement metrics from distributed respondents worldwide.
The project was undertaken by a team of four students:
- Danilo Brandão - Data Science and Engineering (Team Lead)
- António Almeida - Physics Engineering
- Clara Costa - Psychology
- Renata Fontes - Biochemistry
To analyze the correlation between MindProber's metrics and sports events, the team followed these steps:
-
Data Collection:
- Obtained data from the 2022 FIFA World Cup provided by the Statsbomb platform, which tracks various metrics from sports matches.
- Extracted relevant events from the data files using Python and Pandas, exporting each data frame to an Excel file.
- Manually synchronized event times with video feeds on the MindProber database.
-
Metrics Calculation:
- Uploaded the synchronized spreadsheets to the MindProber platform to calculate audience impact score values for each event.
- Conducted an exploratory data analysis to extract insights and information.
-
Analysis and Visualization:
- Explored the correlation between goal probability and audience response, identifying a positive relationship.
- Investigated which players generated higher audience response during matches.
- Utilized event coordinates to determine high-impact zones on the field for different event types.
- Created visualizations, including heatmaps, shot maps, and time-series representations, to provide meaningful insights.
The project yielded impactful results, highlighting the correlation between MindProber's emotional engagement metrics and sports events. The team discovered insights such as the relationship between goal probability and audience response, influential players, and high-impact zones on the field. These findings helped drive MindProber's decision-making process and provided suggestions for incorporating additional value into their existing services.
This project offered a valuable opportunity to apply data science skills to a real-world problem, using messy and diverse data. By addressing MindProber's specific challenge, the team contributed to improving the company's services and offered insights into potential investments in external data sources. This README provides an overview of the project, but more detailed information, code, and visualizations can be found in the repository's files.
Please feel free to explore the repository for a deeper understanding of the project and its outcomes. If you have any questions or feedback, we would be delighted to hear from you.
Thank you for your interest!
Danilo Brandão