B.tech electrical student at iit delhi, interested in football analytics, tracking data, and machine learning. I'm competing in More Than a Hack 2027 (Mobile World Capital Barcelona × Barça Innovation Hub × Metrica Sports).
📌 I'm early in my journey and building in public. This profile shows what I'm using now, what I'm learning, and what I'm building next.
- Estimating team possession per frame from player and ball tracking data
- Visualizing football data on the pitch
- Learning the football-analytics toolchain (kloppy, mplsoccer, floodlight)
Languages: Python Data: pandas, NumPy Visualization: matplotlib Tools: Git, GitHub, Jupyter
- scikit-learn: classification, cross-validation, feature engineering
- Gradient boosting: LightGBM, XGBoost
- Time-series smoothing: rolling filters, hidden Markov models (hmmlearn)
- Football data libraries: kloppy, mplsoccer, floodlight
- SQL
- Streamlit for turning analysis into small apps
- Git and GitHub basics
- Python for data analysis (pandas, NumPy)
- Data visualization (matplotlib, Plotly, mplsoccer)
- ML fundamentals (scikit-learn)
- Gradient boosting (LightGBM / XGBoost)
- Sequence models and smoothing (HMM, then PyTorch basics)
- Tracking-data concepts: coordinates, frames, homography, possession, pitch control
- Event-data concepts: xG, xT, passing networks, PPDA
- Computer vision basics (OpenCV)
- Docker and reproducible environments
- Dashboards with Streamlit
| Project | What it is | Status |
|---|---|---|
| football-possession-estimator | Per-frame team possession from tracking data | 🚧 In progress |
| football-analytics-learning-log | Notes and notebooks while learning | 🚧 In progress |
| pitch-viz-toolkit | Reusable pitch plots and animations | 📅 Planned |
| statsbomb-open-data-explorer | EDA and visuals on open event data | 📅 Planned |
- More Than a Hack 2027: registered, Initial Challenge in progress (possession estimation)
Open to teaming up for football analytics hackathons.
LinkedIn · Email: saideepojha2033@gmail.com