Build your shortlist with position, league, and season context in one pass.
This surface is optimized for fast analyst triage. Query the player pool, inspect the return set, and move strong candidates directly into the workspace for compare or reporting.
Generate a full scouting packet without leaving the workspace.
Use an exact ID when you know the target, or combine fuzzy name, season, and league when you are still narrowing context. The report below remains unchanged; only the intake surface is being upgraded.
One wall for player lenses, market scans, and league‑level pattern reading.
Seven modular surfaces wired to the player-charts API. Move between individual player reconstruction and database-wide analytics without switching context.
/database/ratings — filter by position, league, season, and score range.Every player score in one surface.
The ratings endpoint returns scored player records with overall rating, positional scores, and season metadata. Use the filters below to narrow the view or explore the full population.
/database/role-cluster-results — GMM-based player segmentation.Data-driven role archetypes for every player.
The clustering endpoint returns GMM-assigned roles with membership probabilities. Players are grouped by statistical similarity rather than nominal position — revealing hidden archetypes that traditional labels miss.
Batch predict an entire cohort from a single file upload.
Upload a player spreadsheet — the model returns ratings, cluster assignments, and a cohort-level decision brief. One file in, full intelligence out.