This project is dedicated to reproducing a near-professional environment for code development. It simulates real-world practices, including:
- Team structure and collaboration
- Workload planning and division using GitLab Issues
- Weekly team meetings for coordination
The subject of our work is the analysis of football matches and the prediction of future game outcomes based on historical data and statistical modeling.
The system supports three main modes of operation:
-
CLI Mode
- User interacts directly with the main application through the command line.
-
Client-Server CLI Mode (TCP/UDP)
- User communicates with a CLI client application.
- The client connects to a server that performs calculations and accesses the data.
-
GUI API Mode
- User interacts with a web-based GUI.
- Hosted on a Tomcat server running a Java EE application.
-
View detailed information on:
- Teams
- Players
- Coaches
- Matches
-
Predict win rate probabilities for upcoming matches using historical data:
- User selects a home team and an away team
- The model calculates probabilities based on:
- Team performance statistics
- Head-to-head history
Note: Predictions are directional. That is, analysis for Team A (home) vs Team B (away) may yield different results than the reverse.
-
Raw data obtained from Kaggle, combining multiple datasets:
- Players
- Teams
- Matches
-
Extensive ETL (Extract, Transform, Load) was performed to:
- Clean the data
- Resolve inconsistencies
- Merge multiple datasets into a single concise database
-
Due to data gaps in some datasets (e.g., lack of women's data), the scope was restricted to men's football only.
The project is available in two build formats:
- Java SE:
.jarexecutable for CLI usage - Java EE:
.wararchive deployed via Tomcat 10
Built and managed using Maven.
- Mikołaj Woźniak
- Jakub Szurmak
- Kacper Sochacki
- Szymon Różycki
- Michał Sondej