Machine Learning
Ballinc
Predicts football match results with machine learning and delivers them through a Telegram bot.
Ballinc
// Problem
Football predictions are usually based on gut feeling. The goal was to see how far a data-driven model could go — and to make its predictions easy to consume.
// Solution
Ballinc processes historical match data with pandas, trains a scikit-learn model to predict outcomes, and sends the predictions to users through a Telegram bot.
// Architecture
- Data pipeline: collecting and cleaning match data with pandas
- Feature engineering + scikit-learn prediction model
- Telegram Bot API as the user interface
// Highlights
- Match data processing
- ML-based prediction model
- Predictions delivered via Telegram bot
// What I learned
- Feature engineering matters more than model choice on noisy sports data
- Shipping an ML model behind a simple, familiar interface