baseball-field-viz

A Python library for drawing baseball fields and spray charts from Statcast data.

Companion to savant-extras — together they provide a seamless MLB analysis pipeline.

Installation

pip install baseball-field-viz

Why this library?

pybaseball's built-in spraychart() doesn't support overlaying heatmaps (e.g. seaborn kdeplot). This library gives you a matplotlib Axes object directly, so you can layer any plot on top.

from baseball_field_viz import draw_field
import seaborn as sns

fig, ax = plt.subplots(figsize=(10, 10))
draw_field(ax)
sns.kdeplot(data=df_t, x="x", y="y", ax=ax, cmap="Reds", fill=True, alpha=0.6)

Quick Start

import matplotlib.pyplot as plt
from baseball_field_viz import draw_field, spraychart, transform_coords

# Spray chart (one-liner)
fig, ax = plt.subplots(figsize=(10, 10))
spraychart(ax, df, color_by="events", title="Ohtani 2025 - Batted Balls")
plt.show()

Functions

transform_coords(df)

Converts Statcast hc_x/hc_y (screen pixel coordinates) to real field coordinates in feet (home plate as origin).

from baseball_field_viz import transform_coords

df_t = transform_coords(df)
# Adds 'x', 'y' columns (feet)
# x = 2.5 * (hc_x - 125.42)   ← left/right (positive = right)
# y = 2.5 * (198.27 - hc_y)   ← depth (positive = outfield)

draw_field(ax, foul_distance=330, outfield_distance=340)

Draw a baseball field on a matplotlib Axes.

from baseball_field_viz import draw_field

fig, ax = plt.subplots(figsize=(10, 10))
draw_field(ax)
# Now overlay any plot: scatter, kdeplot, histplot, etc.

spraychart(ax, df, color_by="events", title=None)

Draw a spray chart in one call. Internally calls transform_coords and draw_field.

from baseball_field_viz import spraychart

fig, ax = plt.subplots(figsize=(10, 10))
spraychart(ax, df, color_by="events", title="Player - Season")
plt.show()

color_by options:

Full Example with Heatmap

from pybaseball import statcast
import duckdb
import matplotlib.pyplot as plt
import seaborn as sns
from baseball_field_viz import draw_field, transform_coords

# Fetch data
df_raw = statcast(start_dt="2025-03-01", end_dt="2025-10-31")
con = duckdb.connect()
df = con.execute("""
    SELECT * FROM df_raw
    WHERE batter = 660271  -- mlbam_id (e.g. Shohei Ohtani = 660271)
      AND hc_x IS NOT NULL AND hc_y IS NOT NULL
      AND game_type = 'R'
""").df()

df_t = transform_coords(df)

# Hits vs outs heatmap
fig, axs = plt.subplots(1, 2, figsize=(16, 8))

hits = df_t[df_t["events"].isin(["home_run", "double", "triple", "single"])]
outs = df_t[~df_t["events"].isin(["home_run", "double", "triple", "single"])]

draw_field(axs[0])
sns.kdeplot(data=hits, x="x", y="y", ax=axs[0], cmap="Reds", fill=True, alpha=0.6)
axs[0].set_xlim(-350, 350); axs[0].set_ylim(-50, 400)
axs[0].set_title("Hits Heatmap")

draw_field(axs[1])
sns.kdeplot(data=outs, x="x", y="y", ax=axs[1], cmap="Blues", fill=True, alpha=0.6)
axs[1].set_xlim(-350, 350); axs[1].set_ylim(-50, 400)
axs[1].set_title("Outs Heatmap")

plt.tight_layout()
plt.show()

Strike Zone (v0.2.0)

draw_strike_zone(ax, sz_top=3.5, sz_bot=1.5, color="black", lw=2)

Draw a strike zone rectangle on a matplotlib Axes (plate_x / plate_z coordinate system).

from baseball_field_viz import draw_strike_zone

fig, ax = plt.subplots(figsize=(6, 6))
draw_strike_zone(ax, sz_top=3.5, sz_bot=1.5)

pitch_zone_chart(ax, df, color_by="pitch_type", sz_top=None, sz_bot=None, title=None)

Plot pitch locations with strike zone overlay. Uses plate_x / plate_z columns (Statcast standard). If sz_top/sz_bot are not provided, the mean of df["sz_top"]/df["sz_bot"] is used automatically.

from pybaseball import statcast
from baseball_field_viz import pitch_zone_chart

df = statcast(start_dt="2025-03-01", end_dt="2025-10-31")
df_pitcher = df[df["pitcher"] == 592789]  # mlbam_id: Yoshinobu Yamamoto

fig, ax = plt.subplots(figsize=(6, 6))
pitch_zone_chart(ax, df_pitcher, color_by="pitch_type", title="Yamamoto 2025 - Pitch Locations")
plt.show()

The strike zone is automatically sized from per-pitch Hawk-Eye measurements (sz_top/sz_bot) in the Statcast data, reflecting each batter's actual stance.

Requirements

• Python 3.9+
• matplotlib >= 3.5
• numpy >= 1.21
• pandas >= 1.3

Related

• savant-extras — Statcast bat tracking data fetcher
• pybaseball — Statcast data access

License

MIT