LoginSign Up
← Back to all sparks
C

cfbfastr

ANALYTICS
Velocity0.0

R package cfbfastr by sportsdataverse — release notes from GitHub.

github.com/sportsdataverse/cfbfastr ↗

College football's open data client hit v2 — and now reports how many API calls you have left.

college-footballsports-analyticsapi-migrationrate-limitsplay-by-playr-package
◆Current state
cfbfastR retrieves college football data — play-by-play, box scores, betting lines, ratings and recruiting — from the CollegeFootballData API, ESPN endpoints and the sportsdataverse data repository. Version 2.0.0 in September 2025 was the first release in over three years and rebuilt the package against CFBD's v2 API. Every load_cfb_*() function changed its underlying source to comply with CFBD's terms, the play-by-play dataset gained team and game identifiers users previously had to join in themselves, and a batch of new endpoints arrived covering opponent-adjusted metrics, FPI ratings and live scoreboard and play data.
◆Where it's heading
The package's direction is now set by the data provider rather than by its own plans, and that provider has moved to metered access — the free tier is capped at 1,000 calls a month, with limits tied to membership level. cfbd_api_key_info() reporting a user's tier and usage is the clearest sign of that shift: quota is now something an analysis has to manage. The long gap before 2.0.0 and its arrival largely through a first-time contributor also indicate a package sustained by community effort rather than steady maintenance.
◆Prediction
The live scoreboard and play endpoints are the natural place for the next work, since they are the ones that benefit from in-season iteration. Given the release notes warn users to check their pipelines, follow-up fixes for the changed loading functions are likely before anything new lands.

◆Recent moves

  1. 11mo ago

    Rebuilt on CFBD API v2 with metered access and live endpoints

    ⚡ SPARK

    The release that ends a three-year silence and resets the package against a data source that has changed commercially as well as technically. Where earlier versions accumulated endpoints, this one re-points every loading function and makes quota a user-facing concern.

    View source ↗
  2. 4y ago

    ESPN endpoints and repo-backed loaders added

    The last release of the 1.x line, broadening the package beyond a single provider with ESPN calendar, schedule and team stats functions plus win probability processing. The repo-backed loaders introduced here are the ones 2.0.0 later re-pointed for terms compliance.

    View source ↗
  3. 4y ago

    All outputs standardised as tibbles with a custom class

    Making every function return a tibble with a consistent class gave the package a predictable output contract, which matters for a data client whose functions are chained into pipelines. Unclear extra columns were dropped from drive data at the same time.

    View source ↗
  4. 4y ago

    CRAN release with option-restoring cleanup

    The CRAN publication, carrying only the hygiene changes CRAN requires — restoring user options on exit and removing the GitHub check functions.

    View source ↗
  5. 4y ago

    Minor fixes to betting and FPI rating functions

    A development release posted three minutes before the CRAN one, covering small fixes to the betting and ESPN FPI ratings functions.

    View source ↗
  6. 4y ago

    ESPN scoreboard and play-by-play access, with argument cleanup

    A bundled entry covering three point releases. The substance is the first ESPN scoreboard and play-by-play functions, alongside removal of miscalculated columns and of default years that let users run unintentionally broad queries.

    View source ↗