First Pitches: Standing Up the Statcast Pipeline

Wed, 08 Jul 2026 00:00:00 GMT

This is the first post through the full pipeline: a Cloud Run Job pulls pitch-level Statcast data daily and appends it to BigQuery, and this page reads straight from that table and renders a chart. Nothing here is a deep analytical claim yet - the point of this post is narrower and more important than that: confirming that data really does flow from the source all the way to a published page, with no manual step in between.

Querying the pipeline’s output

from google.cloud import bigquery

client = bigquery.Client()

query = """
SELECT
  pitch_type,
  COUNT(*) AS n_pitches,
  AVG(release_speed) AS avg_release_speed,
  AVG(release_spin_rate) AS avg_spin_rate
FROM `maydaystats.mlb_statcast.pitches`
WHERE pitch_type IS NOT NULL
GROUP BY pitch_type
ORDER BY n_pitches DESC
"""

df = client.query(query).to_dataframe()
df

	pitch_type	n_pitches	avg_release_speed	avg_spin_rate
0	FF	1479	95.085936	2340.053414
1	SI	832	93.824760	2213.739183
2	SL	593	85.249747	2434.556492
3	CH	461	85.870065	1746.629067
4	ST	433	83.066051	2686.658199
5	FC	386	89.453627	2413.054404
6	CU	283	80.454770	2629.568905
7	FS	148	88.922973	1570.432432
8	KC	124	84.465323	2385.064516
9	FO	22	83.954545	1258.636364
10	FA	20	63.150000	1503.700000
11	SV	13	82.800000	2873.769231
12	CS	1	67.200000	2449.000000

As of this post, the table holds 4795 pitches across 13 distinct pitch types.

Average release speed by pitch type

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(8, 5))
ax.bar(df["pitch_type"], df["avg_release_speed"], color="#2c3e50")
ax.set_xlabel("Pitch type")
ax.set_ylabel("Average release speed (mph)")
ax.set_title("Average Release Speed by Pitch Type")
ax.spines[["top", "right"]].set_visible(False)
plt.tight_layout()
plt.show()

Figure 1: Average release speed by pitch type, all games loaded so far

What this confirms, and what’s next

Every number and every bar above came from a live query against mlb_statcast.pitches, which itself only exists because a Cloud Scheduler job triggers a Cloud Run Job every morning that pulls the previous day’s Statcast data. As more days accumulate, this same query pattern (and others like it) becomes the basis for real analysis rather than pipeline validation.

The pipeline code behind this post lives here.

Note

This post uses Quarto’s frozen execution (freeze: true): the code above only re-runs when I re-render locally with fresh BigQuery access. The deployed site reuses that committed output rather than re-querying BigQuery on every build, which is what lets Cloudflare Pages build this site without needing my GCP credentials.