Event Dates: February 27–28, 2026
Location: Northeastern University / Network Science Institute
Supported by: PySport
This repository outlines our starter code for our shot decision quality metric. We used a variety of different attacking features such as shot angle, trajectory, position, and shot speed to determine the metric. Shot deciison quality (SDQ) can be used to determine what makes a "good" shot, with it being given a quantitative value on our own scale.
This differs from a stat like expected goals (xG) as xG measures the probability of a shot ending up in a goal, whilst SDQ measures the quality of the decision to shoot. Expected Goals helps to answer the question "How likely is this shot to end up in a goal?" On the other hand, SDQ helps to answer "Was taking this shot the right decision?"
| Name | Personal Email |
|---|---|
| Jaden Hu | jadenhu2005@gmail.com |
| Zoran Shamsi | zoranshamsi@gmail.com |
| Herchelle Jadhav | jadhavherchelle@gmail.com |
| Rishik Kelkar | rishikk871@gmail.com |
-
Google Slides:
https://docs.google.com/presentation/d/1K5Xf7x6dqUS2SBo25WHKKB_SDN9cPxS8eFPImQ93QHE/edit?usp=sharing -
GitHub Repo:
https://github.com/jadenh3/SoccerImpectHackathon
- Hackathon instructions: https://sites.google.com/view/neusportsanalytics/hackathon?
- GitHub template repo: https://github.com/northeasternsportsanalytics-droid/SoccerImpectHackathon
- Python
- Impact OPEN Data
- Kloppy, pandas, mplsoccer, matplotlib
Define an interpretable attacking or defensive metric using event data. Create a metric definition, produce a leaderboard comparing players, present a case study, and discuss limitations.
AI was used as a minor helper in the ideation of SDQ, deciding what features to use for feature engineering, and partly in the creation of the dashboard.