Footvision — Football Match Analyzer
Automated football match analysis system that detects and tracks players from video, generates tactical maps, and visualises team performance metrics through an interactive dashboard.
Python 
PyTorch 
YOLOv8 
OpenCV 
Streamlit 
Pandas 
Matplotlib 
NumPy 
Scikit-learn 
Footvision is an end-to-end computer vision pipeline that turns raw match footage into structured tactical intelligence — from raw video to interactive tactical dashboards.
What it does
Player and ball detection runs via YOLOv8 (Ultralytics) in real time, identifying every actor on the pitch and tracking bounding boxes across frames. Trajectories feed into Pandas for aggregation and into Matplotlib for generating tactical maps: heatmaps, pass networks, and formation snapshots. The whole thing is served through a Streamlit interface that lets you explore team statistics and possession data without writing a single line of code.
Technical notes
Post-processing handles camera panning (homography-based pitch registration) and team assignment via jersey colour clustering. The modular pipeline design means each component — detection, tracking, visualisation — can be swapped or extended independently.