jab13/daily-statcast-data

Daily MLB Statcast Data

This dataset contains MLB Statcast data that is automatically updated daily.

Data Fields

The dataset includes comprehensive pitch-by-pitch data with over 80 fields including:

Basic Information

• game_date, game_pk, home_team, away_team
• batter, pitcher, fielder_2 through fielder_9
• balls, strikes, outs_when_up

Pitch Data

• pitch_type, pitch_name
• release_speed, effective_speed
• spin_rate, spin_axis
• release_pos_x, release_pos_y, release_pos_z

Location & Zone

• plate_x, plate_z
• sz_top, sz_bot
• zone

Outcome Data

• events, description, type
• hit_location, bb_type
• launch_speed, launch_angle
• hit_distance_sc

Advanced Metrics

• woba_value, woba_denom
• babip_value, iso_value
• estimated_ba_using_speedangle
• estimated_woba_using_speedangle

Usage

from datasets import load_dataset
import pandas as pd

# Load the dataset
dataset = load_dataset("jab13/daily-statcast-data")
df = dataset['train'].to_pandas()

# Basic analysis
print(f"Total pitches: {len(df):,}")
print(f"Date range: {df['game_date'].min()} to {df['game_date'].max()}")

# Pitch type distribution
pitch_types = df['pitch_type'].value_counts()
print("Most common pitch types:")
print(pitch_types.head())

# Exit velocity analysis
hits = df[df['launch_speed'].notna() & (df['launch_speed'] > 0)]
print(f"Average exit velocity: {hits['launch_speed'].mean():.1f} mph")

Data Pipeline

This dataset is maintained by an automated pipeline that:

1. Fetches daily Statcast data using pybaseball
2. Cleans and standardizes the data format
3. Merges with existing data and removes duplicates
4. Updates the Hugging Face dataset with new information

Data Quality

• Deduplication: Removes duplicate pitches based on game_pk, at_bat_number, and pitch_number
• Standardization: Consistent data types and null value handling
• Validation: Ensures data integrity before upload

Updates

The dataset is automatically updated daily at 6 AM UTC to include the previous day's games. Updates typically include:

• All pitch-by-pitch data from completed games
• Corrected data for any previously uploaded games
• Advanced metrics and calculated fields

License

MIT License - Data is publicly available from MLB Advanced Media.

Attribution

Data sourced from MLB Advanced Media via the pybaseball Python package.