Football Decision Intelligence Lab
Team Prototype
Explore team styles, seasonal context, underlying performance, and explainable points drivers using a local first-build export.
EliOra internal R&D · Not client work · Analytics logic and DOM IDs preserved for prototype continuity.
Debug info
Filters
Team Radar Comparison
Compare selected teams across chosen metrics. Percentiles show where each team sits in the current pool.
Top Teams
Ranks teams by a single metric within the selected seasons and points filter.
Holistic Score Bubble Chart
Bubble size = average z-score across selected composite metrics. Color indicates season when multiple seasons are selected.
Use X/Y to compare results vs process; labels show the top 10 by composite score.
Deep Dive Insights
Executive-friendly views that surface process vs results and team style identity.
| Metric | Leader | Value |
|---|
Insight Controls
Adjust the highlighted teams directly while viewing insights.
Process vs Results
xG Diff vs Points. Quadrants show process strength vs results.
Deserved vs Actual
Process score vs points (z‑scores). Above the line = overperforming process.
Team Style Map
Possession % vs shots on target. Reveals style identity by season.
Similarity
Similarity uses Euclidean distance across core z-score features within the filtered dataset, shown as a relative percentage (higher is more similar).
Best for comparing team profiles, not predicting results.
| Team | Season | Points | Similarity % |
|---|
Team Table
Sortable view of the filtered dataset. Blanks indicate missing or unavailable values.
Current View Report
What Drives Points?
Coach-facing view of the strongest patterns associated with points in historical team data.
Model benchmark summary
Grouped cross‑validation by season. Lower MAE/RMSE is better.
What this means
Process drivers are emphasized for sustainable decision‑making.
Top process drivers
Process model uses non‑leaky performance indicators.
Top descriptive drivers
Descriptive model includes results‑adjacent signals for context.
Decision tree view
Simplified tree showing main split patterns associated with higher-point teams.
Model response curves
Partial dependence plots for the best non‑linear model.
Coach takeaway report
Based on historical team data; use as guidance, not certainty.