MLB Statcast + Bat Tracking Dataset (2024-2025)

kaggle
baseball
mlb
data-analysis
statcast
Author

Yasunori Morishima

Published

February 10, 2026

I’ve published a comprehensive MLB Statcast dataset with Bat Tracking metrics for the 2024-2025 seasons on Kaggle.

Dataset: MLB Statcast + Bat Tracking (2024-2025)

What is Bat Tracking?

Bat Tracking is a revolutionary measurement technology introduced by MLB in 2024 that quantifies batting mechanics using high-speed cameras. It captures:

  • bat_speed: Bat velocity at contact (mph)
  • swing_length: Path length of the bat to contact (feet) - shorter indicates more efficient mechanics
  • swing_path_tilt: Angle of swing plane (degrees) - positive = uppercut, negative = downward
  • attack_angle: Bat angle relative to pitch trajectory (degrees)
  • attack_direction: Horizontal bat angle (degrees) - indicates pull vs opposite field tendency

While traditional Statcast focused on batted ball outcomes (exit velocity, launch angle), Bat Tracking quantifies the swing itself - opening new dimensions for analysis.

Dataset Overview

Basic Information

  • Period: 2024-2025 Regular Seasons
  • Rows: 1,443,802 pitches (~1.4M)
  • Columns: 118 metrics
  • Size: 808MB
  • Bat Tracking Coverage: ~46.6% of pitches (~672,000 swings)

Key Metrics

Bat Tracking Metrics - bat_speed, swing_length, swing_path_tilt - attack_angle, attack_direction - intercept_ball_minus_batter_pos_x/y_inches

Statcast Metrics (Traditional) - release_speed, spin_rate, movement (pfx_x, pfx_z) - launch_speed, launch_angle, hit_distance_sc - estimated_woba, estimated_ba_using_speedangle - pitch_type, zone, description, events

Game Context - game_date, game_type, inning, count (balls/strikes) - player_name, batter, pitcher, stand, p_throws - home/away score, win expectancy

All 118 columns have detailed descriptions on Kaggle.

How to Access

The dataset is easily accessible via pybaseball’s statcast() function. Bat Tracking metrics are automatically included.

from pybaseball import statcast
import pandas as pd

# 2024 season
df_2024 = statcast(start_dt='2024-03-20', end_dt='2024-09-29')

# 2025 season
df_2025 = statcast(start_dt='2025-03-27', end_dt='2025-09-28')

# Combine
df = pd.concat([df_2024, df_2025], ignore_index=True)

# Check bat tracking coverage
print(f"Bat tracking coverage: {df['bat_speed'].notna().sum() / len(df) * 100:.1f}%")

Data generation notebook is available on Google Colab: Generate Dataset

Analysis Notebook

I’ve created a comprehensive analysis notebook demonstrating the dataset’s potential:

MLB Bat Tracking Analysis 2024-2025

Analysis Includes: 1. Bat Tracking coverage analysis 2. Distribution of bat_speed, swing_length, swing_path_tilt 3. Correlation: bat_speed vs launch_speed 4. Top 10 bat speed leaders by player 5. Average bat speed by pitch type 6. Heatmap: attack_angle vs attack_direction

Use Cases

This dataset enables various types of analysis:

1. Swing Mechanics Analysis

  • Efficient swing path (swing_length vs bat_speed)
  • Uppercut effectiveness (swing_path_tilt vs launch_angle)
  • Fastest bat speed rankings

2. Pitch-Swing Relationships

  • Which pitch types generate higher bat_speed?
  • Swing tendencies by count
  • Relationship between pitch location and attack_direction

3. Batted Ball Outcomes

  • bat_speed → launch_speed conversion efficiency
  • swing_length impact on batting average
  • attack_angle correlation with barrel rate

4. Player Development & Scouting

  • Evaluating draft prospects’ swing efficiency
  • Tracking player swing improvements
  • Opponent tendency analysis

Comparison with Other Datasets

My previous baseball datasets:

  1. Japanese MLB Players Statcast (2015-2025)
    • 25 pitchers + 10 batters detailed data
  2. MLB Bat Tracking Leaderboard (2024-2025)
    • Season-aggregated data for 452 players
  3. MLB Pitcher Arsenal Evolution (2020-2025)
    • Pitch mix changes over time

This Dataset’s Unique Features: - Pitch-by-pitch granularity - Most comprehensive dataset including Bat Tracking metrics - Two full seasons of data

Conclusion

Bat Tracking adds a new dimension to baseball analysis. Instead of seeing only results, we can now quantify the process of hitting.

I hope this dataset contributes to baseball analytics, machine learning research, and sports science.