Everything Football / Soccer

Football (soccer) data and analytics projects. Current data flow, schedule and stores: docs/SYSTEM_MAP.md. Older full topology: docs/TOPICAL_MAP.md (predates the FotMob/WhoScored work).

FBref · Understat · Transfermarkt · football-data · ClubElo · FotMob · WhoScored (Opta)
StatsBomb/Wyscout open · Baseball Savant
        │  soccerhub readers (fetch → parquet cache → manifest)
        v
  pipelines: xref → player_season · player_match · shots · VAEP (action_score) · links
        │  GitHub Actions, Mon (plain HTTP/CSV sources)
        │  + local launchd, daily 06:00: one Big-5 league per day, ISL on Tuesday
        v
  Supabase Postgres + Storage  ←  source of truth (RLS: anon read-only)
        │
        v
  site/ dashboards (Netlify)  ·  /api/ask → Render agent

Projects

1. Soccer Data Hub (soccerhub)

Python package: unified fetch layer + pipelines for open football data, into Supabase.

• Big 5: 18 seasons (2008–) of player season stats, market values and transfers, entity-resolved across FBref and Transfermarkt; Understat xG and shots (2014–).
• Per-match player stats from FotMob for the Big 5 and the Indian Super League, 2021/22 →.
• On-ball action values (VAEP) from StatsBomb 2015/16, Wyscout 2017/18 and WhoScored (Opta) 2025/26 → current season.

See Soccer Data Hub/README.md.

2. Site (site/)

Static dashboards on Netlify reading Supabase directly (anon key, select-only). Live: landing page, player dashboard (career values, G+A, transfers, shot map), match centre, team analysis (VAEP action maps with per-match filter and touch heatmap), baseball pitch quality. Live at https://sports.karthikiyer.info (Netlify custom domain). Deployed by Netlify on push to main (netlify.toml, publish dir site/). netlify.toml also rewrites /api/ask to the Render agent — same-origin, no CORS.

3. Baseball Analytics

Pitch-quality pipeline on Statcast. See Baseball Analytics/BASEBALL.md.

4. Player Performance Analysis (legacy, frozen)

FIFA player-scouting toolkit, 2015–2022 datasets: ETL, Dash dashboard, statistical EDA. Standalone — shares no code with soccerhub. See docs/product/football-scout-machine.md.

---

Copyright © 2024–2026 Karthik Sivaraman Iyer. All rights reserved.