NFL Player Performance
In DevelopmentA daily-updating NFL data pipeline that ranks players by position, built on AWS with Docker.
Started September 2026 · Updated September 2026
What It Does
A data engineering project that pulls current-season NFL stats every day, cleans and models them, and publishes position rankings (QB, RB, WR, TE, K) to a website. Each position gets two ranking views side by side: a composite performance score and PPR fantasy points, with weekly snapshots to show rank movement. The architecture and ranking method are designed; building the pipeline is the next step. A mobile app is planned as a second phase.
Key Features
- Daily ingestion of NFL stats from the open nflverse datasets
- Position rankings using a composite score of efficiency and production, plus a PPR fantasy view
- Weekly rank snapshots to show players moving up or down
- Data quality tests on every run
- Backtest of the ranking method on past seasons, with a public methodology page
Technologies
- Python
- Docker
- AWS
- Terraform
- dbt
- DuckDB
- Athena
- S3
Architecture
A scheduled container task on AWS (EventBridge and ECS Fargate) ingests raw data into an S3 data lake, dbt models it into staging and mart layers, and a publish step writes rankings as JSON behind CloudFront for the website. Infrastructure is defined in Terraform, and the same Docker image runs locally and in the cloud.
Future Plans
Add defensive player rankings, and build a mobile app on the same published data.