Measuring the invisible force — quantifying how much defensive attention a player commands without touching the ball.
Interactive Power BI dashboard — filter by Player, Team, and Matchup to explore gravity and elasticity across the league.
Recording.2026-03-06.172616.mp4
The best scorers in basketball don't just score — they warp the defense around them, creating open looks for teammates by simply being on the floor. This is gravity: a player's ability to pull defenders away from their assignments.
The Defensive Gravity Model quantifies this off-ball influence using Euclidean spatial analysis. It also introduces Elasticity — measuring how much a player's efficiency fluctuates based on defensive pressure — to reveal who stays consistent under duress and who wilts.
- X-axis: Average Gravity — the spatial force a player exerts on opposing defenders (measured in Euclidean defender displacement units)
- Y-axis: Average Elasticity — how much a player's performance varies under different levels of defensive pressure
- Bubble size: Shot volume
- High Gravity + Low Elasticity = elite, pressure-resistant scorer
- High Gravity + High Elasticity = streaky, pressure-dependent scorer
- Spatial density map of a player's shot origins overlaid on the half-court
- Bright red core = highest shot concentration zones
- Reveals where defenders must respect a player's threat
- A simplified court zone visual showing the player's primary area of operation
- Green fill = dominant zone, cross marker = centroid of shot activity
- A single numerical gravity score (displayed in yellow) representing the total defensive displacement a player generates
- Higher score = more defenders pulled further from their assignments
- Full court scatter of every shot attempt with directional lines from origin point
- Yellow dots/lines show shot trajectory patterns — reveals pull-up range, drive angles, and post-up tendencies
| Tool | Usage |
|---|---|
| Power BI | Dashboard, spatial visualizations, DAX gravity scoring |
| Python (NumPy, SciPy) | Euclidean distance calculations, KDE spatial modeling |
| NBA Stats API / Tracking Data | Defender positioning, shot coordinate data |
| DAX | Gravity Score aggregation, elasticity variance calculations |
LeBron James and Kevin Durant both post high Gravity Scores (10K+ range) but differ significantly in Elasticity — LeBron maintains near-zero elasticity variance under pressure, while high-volume perimeter shooters show much wider elasticity swings. This confirms that gravity without consistency is a defensive luxury, not a nightmare.
Gravity Score = Σ (Defender_Displacement × Shot_Threat_Weight × Court_Zone_Multiplier)
Where:
Defender_Displacement = Euclidean distance defender moves from their assignment
Shot_Threat_Weight = based on historical FG% from that zone
Court_Zone_Multiplier = 3PT zones weighted higher than mid-range
- Clone the repo and open the
.pbixfile in Power BI Desktop - Use Player Name, Team Name, and Matchup filters to explore any player
- The Elasticity chart plots all players league-wide — use it to find outliers
- Select a specific player to isolate their Heat Map, Shot Zone, and Gravity Score
Defensive-Gravity/
├── data/
│ └── tracking_data_cleaned.csv
├── notebooks/
│ └── gravity_model.ipynb
├── dashboard/
│ └── DefensiveGravity.pbix
└── README.md
| Part | Project | Focus |
|---|---|---|
| 1 | Spatial Efficiency Engine | Shot quality & creation mapping |
| 2 | Momentum Volatility Index | Game-flow & run detection |
| 3 | Defensive Gravity Model (you are here) | Off-ball spatial influence |
| 4 | Behavioral Archetyping | Player segmentation & style |