Analysts and researchers
Every NFL play since 1999, already carrying EPA
1.2 million plays with expected points and win probability, plus Next Gen Stats tracking in the same database.
Football research usually means stitching together play-by-play files and Next Gen Stats, each with its own player identifiers and its own idea of what a week is.
The join is the work. The analysis is what you wanted to do.
What to build
EPA by situation
Pull plays filtered by down, distance and field position with EPA already attached, and build the situational splits without computing the model yourself.
Player efficiency beyond the box score
CPOE, air yards, separation and pressure rate come from Next Gen Stats and Pro Football Reference, keyed to the same players as the play-by-play.
Situational tendencies
Filter the play data by down, distance, red zone, play type and win-probability band, so tendency questions do not need a separate vendor.
What it will not do
- Next Gen Stats begins in 2016. Earlier seasons have the plays without those columns.
- Tracking data is the published aggregates, not the raw player-tracking coordinates.
Worth reading first
Start with 500 free credits every month
No card. REST or MCP.