How psubhajit042-tech/ipl-winodds-engine works
This repository is an IPL chasing-team win-probability pipeline. Its main workflow loads match and delivery data, builds second-innings features, trains and evaluates match-grouped classifiers, then uses a trained model for live predictions and historical over-by-over analysis; visualization renders the resulting metrics and progressions. The README also documents an interactive predictor. Source excerpts establish the main orchestration and key data transfers; some training internals and README-level capabilities are represented at module level.
Main entry points: main.py Pipeline runner · model_training.py Grouped training · data_preprocessing.py Feature engineering · predictor.py Live prediction
Components 14
Pipeline orchestration
- Pipeline runner
main.py
Match data
- Match records
- Ball deliveries
- Feature engineering
data_preprocessing.py - Training dataset
Model training
- Grouped training
model_training.py - Classifier pipelines
model_training.py - Saved models
Prediction and charts
- Live prediction
predictor.py - Match progression
predictor.py - Metrics chart
visualization.py - Progression chart
visualization.py
Other components
- Operator
- Prediction results
Connections 20
- Operator → Pipeline runner: runs
main.py - Pipeline runner → Feature engineering: calls
main.py - Match records → Feature engineering: loads
data_preprocessing.py - Ball deliveries → Feature engineering: loads
data_preprocessing.py - Feature engineering → Pipeline runner: returns features
main.py - Pipeline runner → Training dataset: writes
main.py - Training dataset → Grouped training: reads
model_training.py - Pipeline runner → Grouped training: calls
main.py - Grouped training → Classifier pipelines: fits and evaluates
model_training.py - Grouped training → Pipeline runner: returns metrics and models
main.py - Pipeline runner → Saved models: saves
main.py - Saved models → Live prediction: loads model
predictor.py - Pipeline runner → Live prediction: requests prediction
main.py - Pipeline runner → Match progression: requests analysis
main.py - Feature engineering → Match progression: returns delivery details
main.py - Live prediction → Prediction results: returns probabilities
predictor.py - Match progression → Progression chart: provides progression data
main.py - Pipeline runner → Metrics chart: plots metrics
main.py - Pipeline runner → Progression chart: plots progression
main.py - Grouped training → Metrics chart: returns plot metrics
main.py