You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

This dataset requires a paid license. Please purchase access at the link below and provide your Hugging Face username. Once purchased, submit an access request on this page. Access is granted manually.

Log in or Sign Up to review the conditions and access this dataset content.

MLB Weather and Ballpark Run Environment Dataset for Total Research

Question this dataset helps answer: Which MLB weather and ballpark contexts historically align with higher or lower run environments before building totals, DFS, or fantasy research workflows?

This dataset helps totals researchers, fantasy baseball analysts, and sportsbook-market researchers compare MLB run environments using first-pitch weather, ballpark dimensions, roof context, field-oriented wind components, local start time, and rolling team scoring context.

Use the public sample to check:

  • Which parks, roof types, and weather-station matches are represented in the preview
  • Which columns are pregame weather or rolling-context features versus postgame scoring outcomes
  • Which temperature, wind, and run-environment combinations appear in the sample before buying the full file

Preview vs full dataset:

  • Preview: 1200 rows and 84 columns
  • Full: 24005 rows and 84 columns

Full dataset: https://thearticulated.gumroad.com/l/ubegdh

This product is built around a recurring workflow problem that sits between sports betting research, fantasy baseball, and dashboard analysis. Daily MLB weather pages are widely used, but most of them are ephemeral. They help with today's slate, not with historical comparisons across seasons, parks, roof types, wind directions, and scoring environments. Buyers who want to answer questions about totals or run context usually have to rebuild the same stack from scratch: MLB schedules, venue metadata, roof status, local start times, weather observations, and team trend features. This dataset packages that structure directly into a single reproducible table.

Each row represents one completed MLB regular-season game. The coverage runs from April 3, 2016 through June 10, 2026 and includes 24005 game rows across 84 columns. That row count matters. It is large enough for filtering by team, park, month, roof type, first-pitch weather regime, and rolling team form while still retaining meaningful sample sizes. Buyers can use the file for historical dashboards, weather segmentation, park-level comparisons, or time-aware modeling without first spending hours rebuilding raw sources.

The pregame side of the table is built around the weather-and-context problem that buyers already care about. Key feature groups include venue identity, park dimensions, roof type, local start hour, nearest-station weather provenance, first-pitch temperature, humidity, precipitation, wind speed, wind gusts, wind direction, field-oriented wind-out and crosswind components, barometric pressure, cloud cover, team rest gaps, and rolling team offense or defense features. That combination is the unique angle. A generic baseball game log might show only teams and scores. This file adds the context buyers actually search for when they want to compare run environments across parks and weather setups.

The ballpark orientation piece makes the dataset more useful than a raw weather merge. Instead of stopping at wind direction alone, the build converts the weather-station wind reading into a park-aware component that estimates whether the wind was blowing out toward center field or across the field. That gives buyers a better research surface for questions like whether warm temperatures with outward wind at specific parks line up with higher scoring environments more often than cooler or inward-wind setups. It also makes the sample easier to understand because the park-and-wind relationship is explicit in the schema rather than hidden in a manual notebook transformation.

Rolling team features help separate short-term form from environment. The dataset includes pregame rolling averages for runs scored, runs allowed, total runs, hits, and win rate, plus season-to-date runs scored and allowed averages for both home and away teams. Those fields make the dataset useful beyond weather-only filtering. A buyer can study whether a high-scoring outcome came from a hot offense entering a favorable weather window, a struggling pitching environment, or a combination of both. That is closer to the real workflow than simply sorting games by temperature.

The postgame side of the table stores the actual results. Those result columns include home and away score, total runs, run differential, hits, errors, and home-team win outcome. That separation matters. It keeps the pre-event columns usable as responsible research features while preserving the actual game outcome as a target for comparison. The schema therefore supports both descriptive dashboard work and exploratory modeling workflows that need a clean line between information available before first pitch and information only known after the game ended.

Weather coverage is intentionally transparent rather than overstated. Hourly weather is matched through the nearest public station, and the dataset includes weather station identifiers, names, and approximate distance from the venue. That provenance matters when a buyer wants to assess the credibility of a historical weather match at a retractable roof park, a downtown stadium, or an international venue. Some rows may have missing weather because the nearest public station did not provide a clean hourly match at the exact rounded first-pitch time, but the row still retains useful park and team context. That is a better tradeoff than filling historical gaps with invented values.

Responsible use note

This dataset is for research, analysis, dashboard building, educational use, and exploratory modeling. It does not provide betting advice, picks, guarantees, or predictions of future outcomes. It should not be marketed as a shortcut to guaranteed profit, and it should not be interpreted as proof that any single schedule angle automatically creates an edge.

Modeling notes

  • Each row represents one completed MLB regular-season game.
  • Pregame features include park geometry, roof type, local start hour, nearest-station hourly weather, rest gaps, and rolling team scoring context.
  • Postgame outcomes include runs, hits, errors, run differential, and home-team win result.
  • The coverage window runs from April 3, 2016 through June 10, 2026.
  • The full dataset contains 24005 rows, which is large enough for train-test splits by season, by month, by park, or by pregame weather regime.
  • A sensible modeling workflow would split by season or by date rather than shuffling randomly, because the product is time-ordered and season context changes.
  • Weather is station-based rather than stadium-sensor-based, so buyers should treat the station-distance fields as part of data-quality assessment.
  • Retractable-roof and dome parks remain useful analytical rows because the roof-type fields are explicit and filterable.
  • The dataset is descriptive infrastructure, not a finished predictive system.

Why the free sample matters

The public sample is not filler. It proves that the buyer task is real. A buyer can open the sample and immediately verify that the file contains first-pitch weather fields, park geometry, roof metadata, station provenance, rolling offense and defense context, and clear result columns. They can see whether the product answers the practical question they already have: how weather and ballpark context line up with historical scoring environments. That is exactly what a public discovery asset should do.

Source and build notes

The build uses public MLB Stats API schedule and venue metadata plus public Meteostat station data, then engineers park-aware wind components and rolling team run-environment features locally. The process is reproducible, versionable, and does not depend on a private odds feed. The output is a single analysis-ready table rather than a loose collection of raw weather and schedule files that still need substantial cleaning and joins.


Dataset Preview

record_id game_pk season official_date game_datetime_utc game_month game_weekday game_local_start_hour day_night scheduled_innings doubleheader_flag games_in_series series_game_number series_description venue_id venue_name venue_city venue_state venue_country venue_timezone venue_elevation_feet field_azimuth_to_center_deg roof_type turf_type left_line_feet left_center_feet center_field_feet right_center_feet right_line_feet is_dome is_retractable_roof is_open_air home_team_id home_team_abbreviation home_team_name away_team_id away_team_abbreviation away_team_name weather_station_id weather_station_name weather_station_distance_m temperature_f relative_humidity_pct precipitation_mm wind_speed_mph wind_gust_mph wind_direction_from_deg wind_out_to_center_mph crosswind_mph barometric_pressure_hpa cloud_cover_pct home_prior_games_played away_prior_games_played home_days_since_prev_game away_days_since_prev_game rest_advantage_days home_rolling_runs_scored_last_5 away_rolling_runs_scored_last_5 home_rolling_runs_allowed_last_5 away_rolling_runs_allowed_last_5 home_rolling_total_runs_last_5 away_rolling_total_runs_last_5 home_rolling_hits_last_5 away_rolling_hits_last_5 home_rolling_win_pct_last_5 away_rolling_win_pct_last_5 home_season_to_date_runs_scored_avg away_season_to_date_runs_scored_avg home_season_to_date_runs_allowed_avg away_season_to_date_runs_allowed_avg home_score away_score total_runs run_differential home_team_won home_hits away_hits total_hits home_errors away_errors total_errors source_url source_domain last_collected_at
446877 446877 2016 2016-04-03 00:00:00 2016-04-03 17:05:00 4 Sunday 13 day 9 False 3 1 Regular Season 31 PNC Park Pittsburgh PA USA America/New_York 780 116 Open Grass 325 410 399 375 320 False False True 134 Pittsburgh Pirates 138 St. Louis Cardinals 72520 Greater Pittsburgh International 6554.6 35.06 42 0.0 13.857 270.0 12.454 -6.074 1018.5 0 0 4 1 5 3 True 9 5 14 1 0 1 https://statsapi.mlb.com/api/v1.1/game/446877/feed/live statsapi.mlb.com, meteostat.net 2026-06-11 19:41:03
446911 446911 2016 2016-04-03 00:00:00 2016-04-03 20:05:00 4 Sunday 16 day 9 False 4 1 Regular Season 12 Tropicana Field St. Petersburg FL USA America/New_York 15 359 Dome Artificial Turf 315 410 404 404 322 True False False 139 Tampa Bay Rays 141 Toronto Blue Jays KSPG0 Saint Petersburg 2526.6 71.96 33 0.0 6.959 120.0 3.584 -5.965 1017.2 0.0 0 0 3 5 8 -2 False 7 7 14 1 2 3 https://statsapi.mlb.com/api/v1.1/game/446911/feed/live statsapi.mlb.com, meteostat.net 2026-06-11 19:41:03
446873 446873 2016 2016-04-03 00:00:00 2016-04-04 00:37:00 4 Sunday 19 night 9 False 2 1 Regular Season 7 Kauffman Stadium Kansas City MO USA America/Chicago 856 46 Open Grass 330 385 410 385 330 False False True 118 Kansas City Royals 121 New York Mets KMKC0 Kansas City / Harlem 12546.5 75.02 18 0.0 12.738 210.0 12.245 -3.511 1008.2 0.0 0 0 4 3 7 1 True 9 7 16 0 1 1 https://statsapi.mlb.com/api/v1.1/game/446873/feed/live statsapi.mlb.com, meteostat.net 2026-06-11 19:41:03
446875 446875 2016 2016-04-04 00:00:00 2016-04-04 18:10:00 4 Monday 13 day 9 False 3 1 Regular Season 32 Miller Park Milwaukee WI USA America/Chicago 597 129 Retractable Grass 344 371 400 374 345 False True False 158 Milwaukee Brewers 137 San Francisco Giants 72640 Gen Mitchell International Airport 9793.4 33.98 56 17.212 20.0 5.604 16.274 1023.8 6.0 0 0 3 12 15 -9 False 8 15 23 1 0 1 https://statsapi.mlb.com/api/v1.1/game/446875/feed/live statsapi.mlb.com, meteostat.net 2026-06-11 19:41:03
446872 446872 2016 2016-04-04 00:00:00 2016-04-04 19:05:00 4 Monday 15 day 9 False 3 1 Regular Season 2 Oriole Park at Camden Yards Baltimore MD USA America/New_York 33 31 Open Grass 333 410 400 373 318 False False True 110 Baltimore Orioles 142 Minnesota Twins KDMH0 Baltimore 432.8 71.06 26 0.0 1007.2 0 0 3 2 5 1 True 10 7 17 0 0 0 https://statsapi.mlb.com/api/v1.1/game/446872/feed/live statsapi.mlb.com, meteostat.net 2026-06-11 19:41:03

Access Requirements (Paid Dataset)

This dataset is behind manual gated access.

To obtain access:

  1. Full dataset access:
    https://thearticulated.gumroad.com/l/ubegdh

  2. Provide your Hugging Face username at checkout.

  3. Return to this Hugging Face page and click:
    "Request Access"

  4. Your access will be approved within 1-12 hours.

Once approved, you can use the Python snippet at the bottom of this README to load the dataset.


Dataset Structure

Total rows: 24,005

Total columns: 84

Splits

  • data: 24,005 rows

Data Files

  • data: data/data.parquet

Data Dictionary

The table below describes the columns included in this dataset.

column pandas_dtype dataset_type description
record_id str integer-like string Stored as text, but values appear to represent whole numbers. Stable unique identifier for one completed MLB game row.
game_pk str integer-like string Stored as text, but values appear to represent whole numbers. MLB Stats API game identifier stored as text.
season Int64 integer Whole-number numeric column. MLB season year attached to the regular-season game.
official_date datetime64[us] datetime Date or timestamp column. Official calendar date of the game.
game_datetime_utc datetime64[us] datetime Date or timestamp column. Scheduled game start time in UTC.
game_month Int64 integer Whole-number numeric column. Calendar month of the game date.
game_weekday str string Text column. Weekday label for the official game date.
game_local_start_hour Int64 integer Whole-number numeric column. Local scheduled start hour at the venue timezone.
day_night str string Text column. Source label for day or night start context.
scheduled_innings Int64 integer Whole-number numeric column. Scheduled innings count from the game metadata.
doubleheader_flag bool boolean True/false column. Boolean flag showing whether the game was part of a doubleheader sequence.
games_in_series Int64 integer Whole-number numeric column. Number of games in the series when provided by the source.
series_game_number Int64 integer Whole-number numeric column. Sequence number of the game within its series when provided by the source.
series_description str string Text column. Series descriptor such as regular season or opening series.
venue_id Int64 integer Whole-number numeric column. Venue identifier from the public MLB source.
venue_name str string Text column. Venue name for the game.
venue_city str string Text column. Venue city from the public source.
venue_state str string Text column. Venue state or province abbreviation when available.
venue_country str string Text column. Venue country from the public source.
venue_timezone str string Text column. IANA timezone name for the venue.
venue_elevation_feet float64 float Decimal numeric column. Venue elevation from the public source metadata, generally reported in feet.
field_azimuth_to_center_deg float64 float Decimal numeric column. Field azimuth angle toward center field used to derive wind-out and crosswind components.
roof_type str string Text column. Roof type from venue metadata, such as Open, Dome, or Retractable.
turf_type str string Text column. Playing surface or turf type from venue metadata.
left_line_feet Int64 integer Whole-number numeric column. Listed distance down the left-field line.
left_center_feet Int64 integer Whole-number numeric column. Listed distance to left-center field.
center_field_feet Int64 integer Whole-number numeric column. Listed distance to straightaway center field.
right_center_feet Int64 integer Whole-number numeric column. Listed distance to right-center field.
right_line_feet Int64 integer Whole-number numeric column. Listed distance down the right-field line.
is_dome bool boolean True/false column. Boolean flag for dome roof types.
is_retractable_roof bool boolean True/false column. Boolean flag for retractable-roof venues.
is_open_air bool boolean True/false column. Boolean flag for open-air venues where outside weather is more directly relevant.
home_team_id Int64 integer Whole-number numeric column. Home team identifier from the MLB source.
home_team_abbreviation str string Text column. Home team abbreviation.
home_team_name str string Text column. Home team full display name.
away_team_id Int64 integer Whole-number numeric column. Away team identifier from the MLB source.
away_team_abbreviation str string Text column. Away team abbreviation.
away_team_name str string Text column. Away team full display name.
weather_station_id str string Text column. Nearest Meteostat station identifier used to source hourly weather for the venue.
weather_station_name str string Text column. Nearest Meteostat station name used for the weather match.
weather_station_distance_m float64 float Decimal numeric column. Approximate straight-line distance in meters from the venue to the matched weather station.
temperature_f float64 float Decimal numeric column. Archived hourly air temperature near first pitch converted to degrees Fahrenheit.
relative_humidity_pct float64 float Decimal numeric column. Archived hourly relative humidity percentage near first pitch.
precipitation_mm float64 float Decimal numeric column. Archived hourly precipitation amount near first pitch in millimeters.
wind_speed_mph float64 float Decimal numeric column. Archived hourly wind speed near first pitch converted to miles per hour.
wind_gust_mph float64 float Decimal numeric column. Archived hourly peak wind gust near first pitch converted to miles per hour when available.
wind_direction_from_deg float64 float Decimal numeric column. Archived hourly meteorological wind direction in degrees, representing where the wind came from.
wind_out_to_center_mph float64 float Decimal numeric column. Derived wind component in miles per hour toward center field. Positive values imply wind blowing out, negative values imply wind blowing in.
crosswind_mph float64 float Decimal numeric column. Derived cross-field wind component in miles per hour based on venue orientation.
barometric_pressure_hpa float64 float Decimal numeric column. Archived hourly mean sea level pressure near first pitch in hectopascals.
cloud_cover_pct float64 float Decimal numeric column. Archived hourly cloud-cover percentage near first pitch when available.
home_prior_games_played Int64 integer Whole-number numeric column. Number of prior completed games for the home team before this game.
away_prior_games_played Int64 integer Whole-number numeric column. Number of prior completed games for the away team before this game.
home_days_since_prev_game Int64 integer Whole-number numeric column. Calendar days since the home team's prior completed game.
away_days_since_prev_game Int64 integer Whole-number numeric column. Calendar days since the away team's prior completed game.
rest_advantage_days Int64 integer Whole-number numeric column. Home-team rest gap minus away-team rest gap.
home_rolling_runs_scored_last_5 float64 float Decimal numeric column. Pregame rolling average runs scored across the home team's previous five completed games.
away_rolling_runs_scored_last_5 float64 float Decimal numeric column. Pregame rolling average runs scored across the away team's previous five completed games.
home_rolling_runs_allowed_last_5 float64 float Decimal numeric column. Pregame rolling average runs allowed across the home team's previous five completed games.
away_rolling_runs_allowed_last_5 float64 float Decimal numeric column. Pregame rolling average runs allowed across the away team's previous five completed games.
home_rolling_total_runs_last_5 float64 float Decimal numeric column. Pregame rolling average total runs across the home team's previous five completed games.
away_rolling_total_runs_last_5 float64 float Decimal numeric column. Pregame rolling average total runs across the away team's previous five completed games.
home_rolling_hits_last_5 float64 float Decimal numeric column. Pregame rolling average hits recorded by the home team across its previous five completed games.
away_rolling_hits_last_5 float64 float Decimal numeric column. Pregame rolling average hits recorded by the away team across its previous five completed games.
home_rolling_win_pct_last_5 float64 float Decimal numeric column. Pregame rolling win percentage across the home team's previous five completed games.
away_rolling_win_pct_last_5 float64 float Decimal numeric column. Pregame rolling win percentage across the away team's previous five completed games.
home_season_to_date_runs_scored_avg float64 float Decimal numeric column. Pregame season-to-date average runs scored by the home team before this game.
away_season_to_date_runs_scored_avg float64 float Decimal numeric column. Pregame season-to-date average runs scored by the away team before this game.
home_season_to_date_runs_allowed_avg float64 float Decimal numeric column. Pregame season-to-date average runs allowed by the home team before this game.
away_season_to_date_runs_allowed_avg float64 float Decimal numeric column. Pregame season-to-date average runs allowed by the away team before this game.
home_score Int64 integer Whole-number numeric column. Final runs scored by the home team.
away_score Int64 integer Whole-number numeric column. Final runs scored by the away team.
total_runs Int64 integer Whole-number numeric column. Combined final runs scored in the game.
run_differential Int64 integer Whole-number numeric column. Home-team run differential for the game.
home_team_won bool boolean True/false column. Boolean flag showing whether the home team won the game.
home_hits Int64 integer Whole-number numeric column. Final hits recorded by the home team.
away_hits Int64 integer Whole-number numeric column. Final hits recorded by the away team.
total_hits Int64 integer Whole-number numeric column. Combined final hits recorded by both teams.
home_errors Int64 integer Whole-number numeric column. Final errors charged to the home team.
away_errors Int64 integer Whole-number numeric column. Final errors charged to the away team.
total_errors Int64 integer Whole-number numeric column. Combined final errors charged to both teams.
source_url str string Text column. Game endpoint URL from the public MLB Stats API.
source_domain str string Text column. Source domain summary for the dataset row.
last_collected_at datetime64[us] datetime Date or timestamp column. UTC timestamp when this dataset build collected and transformed the row.

Intended Use

This dataset is intended for research, experimentation, analysis, and model prototyping.

Loading the Dataset


import os
from datasets import load_dataset

HUGGINGFACE_API_KEY_KARMANE = os.environ.get("HUGGINGFACE_API_KEY_KARMANE")

dataset = load_dataset(
    "Karmane/mlb-weather-run-environment-research",
    token=HUGGINGFACE_API_KEY_KARMANE,
)

print(dataset)
print(dataset[list(dataset.keys())[0]][0])

# getting the DataFrame itself
# df = dataset[list(dataset.keys())[0]].to_pandas()

Karmane. (2025). MLB Weather and Ballpark Run Environment Dataset for Total Research. Hugging Face. /datasets/Karmane/mlb-weather-run-environment-research

Downloads last month
16