Baseball Savant Statcast Pitch Data Scraper

Extracts pitch-level Statcast data from Baseball Savant (baseballsavant.mlb.com). One row per tracked pitch event: exit velocity, launch angle, xwOBA, barrel rate, spin rate, and more. Optionally also pulls the expected-stats leaderboard for per-player season aggregates.

This is the dataset Python analysts currently pull via pybaseball. The scraper handles the large CSV downloads, chunking, and field coercion so you don't have to.

What It Does

• Fetches Statcast pitch-level data from the statcast_search CSV endpoint
• Covers batter data, pitcher data, or both — configurable per run
• Filters by team abbreviation (e.g., NYY, LAD) to keep individual requests manageable
• Optionally pulls the expected-stats leaderboard (xBA, xwOBA per player, season aggregates)
• Returns clean, typed JSON with consistent field names

Use Cases

• Sabermetric analysts building pitch-mix models, exit velocity distributions, or barrel-rate dashboards
• Fantasy/DFS players tracking real-time statcast trends to identify undervalued hitters or pitchers
• Sports betting models that use xwOBA and spin rate as input features
• Researchers studying pitch design, swing mechanics, or predictive batting metrics
• Data engineers building baseball datasets without managing throttling or CSV parsing

How It Works

1. Input your target season, player type (batter/pitcher/both), and optionally a team abbreviation
2. The scraper fetches the Baseball Savant CSV endpoint with a browser user agent
3. Parses the CSV row by row (RFC 4180-compliant, handles BOM and quoted fields)
4. Coerces all numeric fields and returns one record per pitch event
5. Optionally fetches the expected-stats leaderboard as a second pass

No proxies required. Baseball Savant is an open, public endpoint run by MLB.

Input

Field	Type	Default	Description
season	integer	2024	MLB season year to scrape
playerType	string	batter	Whether to fetch batter, pitcher, or both
mode	string	statcast	Data mode: statcast (pitch-level), leaderboard (season aggregates), or both
teamAbbrev	string	(all teams)	Optional 3-letter team abbreviation to restrict the query (e.g. NYY, LAD, BOS)
minPa	integer	50	Minimum plate appearances for the leaderboard mode
maxItems	integer	15	Maximum records to return. Set to 0 for unlimited (full season = 700k+ rows for all teams)

Example — 2024 NYY batters, statcast mode:

{
  "season": 2024,
  "playerType": "batter",
  "mode": "statcast",
  "teamAbbrev": "NYY",
  "maxItems": 1000
}

Example — full league expected stats leaderboard:

{
  "season": 2024,
  "playerType": "batter",
  "mode": "leaderboard",
  "minPa": 300,
  "maxItems": 0
}

Output

Statcast mode

One record per pitch event. Large seasons can return 700,000+ rows for all teams.

{
  "player_name": "Judge, Aaron",
  "batter_id": 592450,
  "pitcher_id": 519242,
  "game_date": "2024-09-28",
  "pitch_type": "FF",
  "release_speed": 95.2,
  "launch_speed": 109.8,
  "launch_angle": 28,
  "estimated_ba": 0.732,
  "estimated_woba": 0.891,
  "barrel": 1,
  "hit_distance": 423,
  "spin_rate": 2241,
  "events": "home_run",
  "description": "hit_into_play",
  "zone": 5,
  "balls": 1,
  "strikes": 1,
  "source": "statcast"
}

Field	Type	Description
player_name	string	Batter name (Last, First)
batter_id	integer	MLBAM batter player ID
pitcher_id	integer	MLBAM pitcher player ID
game_date	string	Date of the game (YYYY-MM-DD)
pitch_type	string	Pitch type abbreviation (FF = four-seam fastball, SL = slider, CH = changeup, etc.)
release_speed	number	Pitch release speed in mph
launch_speed	number	Exit velocity off the bat in mph
launch_angle	number	Launch angle in degrees
estimated_ba	number	Expected batting average (xBA) based on exit velocity and launch angle
estimated_woba	number	Expected weighted on-base average (xwOBA)
barrel	integer	Barrel flag — 1 if the batted ball meets the barrel threshold, 0 otherwise
hit_distance	number	Projected hit distance in feet
spin_rate	number	Pitch spin rate in RPM
events	string	Plate appearance result (single, home_run, strikeout, walk, etc.)
description	string	Pitch result (called_strike, ball, swinging_strike, hit_into_play, etc.)
zone	integer	Strike zone location (1-9 in zone, 11-14 outside)
balls	integer	Ball count at time of pitch
strikes	integer	Strike count at time of pitch
source	string	Always statcast in this mode

Leaderboard mode

One record per player per season. Typically 250-500 rows depending on the minimum plate appearances filter.

{
  "player_name": "Duran, Jarren",
  "batter_id": 680776,
  "pitcher_id": null,
  "game_date": null,
  "pitch_type": null,
  "release_speed": null,
  "launch_speed": null,
  "launch_angle": null,
  "estimated_ba": 0.267,
  "estimated_woba": 0.34,
  "barrel": null,
  "hit_distance": null,
  "spin_rate": null,
  "events": null,
  "description": null,
  "zone": null,
  "balls": null,
  "strikes": null,
  "source": "leaderboard"
}

The leaderboard shares the same schema as statcast mode — pitch-specific fields are null for leaderboard rows. estimated_ba and estimated_woba reflect season aggregates.

🔍 FAQ

How do I scrape statcast data from Baseball Savant? Baseball Savant Statcast Pitch Data Scraper hits the public statcast_search/csv endpoint directly. Set your season, filter by team if needed, and set maxItems to cap the result set for testing.

How much does this actor cost to run? Baseball Savant Statcast Pitch Data Scraper uses pay-per-event pricing at the standard rate. A full team season (~25,000 pitch events) is inexpensive. A full league season (700,000+ rows) costs more — consider filtering by team or adding a maxItems cap if you only need a sample.

What data can I get from Baseball Savant statcast? Baseball Savant Statcast Pitch Data Scraper returns pitch velocity, launch angle, exit velocity, xBA, xwOBA, barrel flag, spin rate, zone, count, and game context — the core Statcast fields analysts use for modern baseball modeling.

Can I filter by team? Set teamAbbrev to any standard 3-letter MLB abbreviation (NYY, LAD, BOS, HOU, etc.) to pull only pitches involving that team's batters or pitchers. Useful when you don't need the full league dataset.

Does this actor need proxies? Baseball Savant Statcast Pitch Data Scraper does not require proxies. The endpoint is public and accessible with a standard browser user agent.

Why Use Baseball Savant Statcast Pitch Data Scraper?

• No setup overhead — pybaseball requires a Python environment, rate limiting awareness, and manual CSV handling. This actor returns structured JSON in one click.
• Team-level filtering — pull a specific team's pitch data instead of waiting for a full-league 17 MB CSV to download.
• Two modes in one — pitch-level statcast detail and per-player expected-stats leaderboard from the same actor with the same output schema.

Need More Features?

Need additional Statcast fields, date-range filtering, or a specific leaderboard type? File an issue or get in touch.