IPL Analytics - 17 Seasons of Cricket Intelligence
End-to-end analysis of IPL ball-by-ball data from 2008 to 2025, producing team performance, player records, and match pattern dashboards.
Open cricket analytics is usually either shallow or locked away.
Cricket analytics publicly available is either surface-level or behind paid services. This project asks what the data says about player performance, team patterns, and match dynamics without an editorial filter.
Granular source
Cricsheet ball-by-ball JSON: every delivery, wicket, and run from IPL Season 1 through Season 18.
Coverage
1,168+ matches and 600+ players tracked across 17 seasons.
Public intelligence
A modern interactive layer over open cricket data, built to answer questions rather than decorate metrics.
Analysis first, presentation second.
The analysis layer was built separately from the content/social layer. DuckDB handles fast in-process querying of large JSON datasets, FastAPI exposes the analytical backend, and React + Tailwind power the dashboard frontend. Every chart is treated as an answer to a specific cricket question.
The architecture keeps insight reusable.
Ball-by-ball over summary data
Delivery-level data enables phase analysis, pressure metrics, and matchup views impossible with match-level aggregates.
DuckDB over Pandas for querying
Querying 17 seasons of JSON directly keeps iteration fast and avoids stale pre-computed tables.
Separate analysis and content layers
The analytical database is not mixed with social media content, so the same source can power future formats.
A live dashboard with deeper public IPL analysis.
The dashboard covers season-by-season team trends, batting and bowling records, match phase analysis, head-to-head player matchups, and venue-level statistics. It is deeper in analytical scope than most publicly available IPL dashboards.
What I learned
The decision to separate analysis from content was the most important one. It means the data work can be reused, republished, and built on top of rather than tied to a single presentation format. Good analysis is infrastructure, not output.