Accessing Baseball Data for Analysis with R

Sources and Methods

Author

David Dalpiaz

Published

October 4, 2022

This post is currently a work-in-progress. Use this resource outside of STAT 430 Baseball Analytics at Illinois with extreme caution.

Background

The sport of baseball has a long history of data collection dating back to the first “box scores” recorded by Henry Chadwick.

Sources

Today, baseball provides a rich source of data for analysis.

Methods

Flat Files

  • Flat File
    • Plain text: txt, csv, tsv
    • Compressed: zip, gzip, zst
  • https://en.m.wikipedia.org/wiki/Plain_text
  • https://en.m.wikipedia.org/wiki/Data_compression

Serialization

Application Programming Interfaces

Databases

Web Scraping

R Pacakges

baseballr

bbd

Structure

  • What is the smallest possible unit of analysis?
  • What is a row of the data?

Scraping Example

library("dplyr")
library("rvest")
url = "https://www.baseball-reference.com/players/m/marisro01.shtml"
maris = html_table(x = read_html(x = url), )[[1]]
maris_mlb = {
  maris |>
    filter(Lg %in% c("AL", "NL")) |>
    mutate(Year = as.integer(Year),
           Age = as.integer(Age))
}
DT::datatable(maris_mlb)

Package Example

library("dplyr")
war_pitcher = bbd::bref_war_daily_batter()
war_batter  = bbd::bref_war_daily_pitcher()
war_pitcher_2022_nl = {
  war_pitcher |>
    filter(year_ID == 2022, lg_ID == "NL") |>
    summarise(war = sum(WAR), .by = c(player_ID, name_common))
}
war_batter_2022_nl = {
  war_batter |>
    filter(year_ID == 2022, lg_ID == "NL") |>
    summarise(war = sum(WAR), .by = c(player_ID, name_common))
}
war_pitcher_2022_al = {
  war_pitcher |>
    filter(year_ID == 2022, lg_ID == "AL") |>
    summarise(war = sum(WAR), .by = c(player_ID, name_common))
}
war_batter_2022_al = {
  war_batter |>
    filter(year_ID == 2022, lg_ID == "AL") |>
    summarise(war = sum(WAR), .by = c(player_ID, name_common))
}
war_2022_nl = rbind(war_pitcher_2022_nl, war_batter_2022_nl)
war_2022_al = rbind(war_pitcher_2022_al, war_batter_2022_al)
war_2022_nl = arrange(war_2022_nl, desc(war))
war_2022_al = arrange(war_2022_al, desc(war))
DT::datatable(war_2022_al)

Do you notice anything odd about this data?

Database Example

2022 Home Run Races

library("dplyr")
library("DBI")
  • TODO: Add examples of listing all tables in the database.
con = DBI::dbConnect(duckdb::duckdb(), dbdir = "data/statcast.db")
statcast = tbl(src = con, "sc")
hr_races_2022 = {
  statcast |>
    filter(batter_name %in% c("Aaron Judge", "Albert Pujols")) |>
    filter(events == "home_run") |>
    select(
      game_date,
      batter_name,
      stand,
      pitcher_name,
      p_throws,
      pitch_name,
      release_speed,
      release_spin_rate,
      hit_distance_sc,
      launch_speed,
      launch_angle
    )
}
show_query(hr_races_2022)
<SQL>
SELECT
  game_date,
  batter_name,
  stand,
  pitcher_name,
  p_throws,
  pitch_name,
  release_speed,
  release_spin_rate,
  hit_distance_sc,
  launch_speed,
  launch_angle
FROM sc
WHERE
  (batter_name IN ('Aaron Judge', 'Albert Pujols')) AND
  (events = 'home_run')
hr_races_2022 = collect(hr_races_2022)
DT::datatable(
  hr_races_2022,
  colnames = c(
    "Date",
    "Hitter",
    "Stands",
    "Pitcher",
    "Throws",
    "Pitch Type",
    "Pitch Velocity",
    "Pitch Spin",
    "Hit Distance",
    "Exit Velocity",
    "Launch Angle"
  )
)
DBI::dbDisconnect(conn = con, shutdown = TRUE)

Lahman

library("dbplyr")
library("dplyr")
library("DBI")
lahman_db = dbplyr::lahman_sqlite()
DBI::dbListTables(lahman_db)
 [1] "AllstarFull"         "Appearances"         "AwardsManagers"     
 [4] "AwardsPlayers"       "AwardsShareManagers" "AwardsSharePlayers" 
 [7] "Batting"             "BattingPost"         "CollegePlaying"     
[10] "Fielding"            "FieldingOF"          "FieldingOFsplit"    
[13] "FieldingPost"        "HallOfFame"          "HomeGames"          
[16] "LahmanData"          "Managers"            "ManagersHalf"       
[19] "Parks"               "People"              "Pitching"           
[22] "PitchingPost"        "Salaries"            "Schools"            
[25] "SeriesPost"          "Teams"               "TeamsFranchises"    
[28] "TeamsHalf"           "sqlite_stat1"        "sqlite_stat4"