l2m analysis

# Load packages
library(tidyverse)
library(vroom)
library(gt)
library(gtExtras)

# read in full l2m data
dat <- vroom("https://raw.githubusercontent.com/atlhawksfanatic/L2M/master/1-tidy/L2M/L2M.csv")

# count total games requiring an L2M for each team in 2023 -- we'll use this later
total_games <- dat %>% 
  filter(season == 2023) %>%
  select(bkref_id, home_team, away_team) %>% 
  distinct() %>% 
  pivot_longer(-bkref_id, values_to = "team") %>% 
  count(team, sort = T) 

# select relevant columns
df <- dat %>% 
  select(game_details, game_date, period, time, comments, call_type, committing, committing_team, disadvantaged_team, decision, season) %>% 
  # remove duplicates %>% 
  distinct() %>% 
  # limit data to 2023 
  filter(season == 2023)

# hard code some ICs/INCs that were labeled as NA for some reason 
df <- df %>% 
  mutate(
    disadvantaged_team = case_when(
      # klay steps out of bounds. wasnt called. disadvantages ORL, favors GSW
      game_details == "Warriors(129) @ Magic(130)" & decision %in% c("INC", "IC") & (is.na(committing_team) | is.na(disadvantaged_team)) ~ "ORL", 
      # shot clock was reset. shouldnt have been. disadvantages PHI, favors HOU 
      game_details == "76ers(123) @ Rockets(132)" & decision %in% c("INC", "IC") & (is.na(committing_team) | is.na(disadvantaged_team))~ "PHI", 
      # mobley travels. wasnt called. disadvantages IND, favors CLE
      game_details == "Pacers(112) @ Cavaliers(118)" & decision %in% c("INC", "IC") & (is.na(committing_team) | is.na(disadvantaged_team)) ~ "IND", 
      # tre jones interferes with the ball and tosses it the official. wasnt called. disadvantages MEM, favors HOU
      game_details == "Spurs(129) @ Grizzlies(135)" & decision %in% c("INC", "IC") & (is.na(committing_team) | is.na(disadvantaged_team)) ~ "MEM", 
      TRUE ~ disadvantaged_team),
    committing_team = case_when(
      # shot clock was reset. shouldnt have been. disadvantages PHI, favors HOU 
      game_details == "76ers(123) @ Rockets(132)" & decision %in% c("INC", "IC") & (is.na(committing_team) | is.na(disadvantaged_team)) ~ "HOU", 
      TRUE ~ committing_team)
    ) 

# count times an INC (incorrect no call) or IC (incorrect call) went in a teams favor
favor <- df %>% 
  mutate(team_favored = case_when(
    decision == "INC" ~ committing_team,
    decision == "IC" ~ disadvantaged_team
  )) %>% 
  group_by(team_favored) %>% 
  summarize(count_favored = n()) %>% 
  filter(!is.na(team_favored)) %>% 
  arrange(desc(count_favored))

# count times INC or IC went in other team's favor
not_favor <- df %>% 
  mutate(team_not_favor = case_when(
    decision == "INC" ~ disadvantaged_team,
    decision == "IC" ~ committing_team
  )) %>%
  group_by(team_not_favor) %>% 
  summarize(count_not_favored = n()) %>% 
  filter(!is.na(team_not_favor)) %>% 
  arrange(desc(count_not_favored))

# combine two bad call data frames into one
bad_calls_df <- full_join(favor, not_favor, by = c("team_favored" = "team_not_favor")) %>% 
  rename("team" = "team_favored") %>% 
  left_join(total_games, by = "team") %>% 
  transmute(Team = team, 
            `L2Ms` = n, 
            `Errors` = count_favored + count_not_favored, 
            `For` = count_favored, 
            `Against` = count_not_favored,
            `Diff.` = count_favored - count_not_favored) %>%
  arrange(desc(`Diff.`)) %>% 
  print(n = 30)

# get the min and max of the differences 
min_max <- max(c(abs(min(bad_calls_df$`Diff.`)), abs(max(bad_calls_df$`Diff.`))))

# create table 
bad_calls_df %>% 
  gt() %>% 
  tab_header(
    title = md("**L2M Reports**"), 
    subtitle = paste0("2022-23 Regular Season | Updated ", format(Sys.Date(), format="%B %d, %Y"))
  )  %>% 
  gt_theme_538() %>% 
  gt_hulk_col_numeric("Diff.", trim = TRUE, domain = c(min_max, min_max*-1)) %>% 
  tab_options(
    heading.title.font.size  = 16,
    heading.title.font.weight = 'bold',
    heading.subtitle.font.size = 11,
    column_labels.font.size = 10.5,
    table.font.size = 10
  )  %>%
  cols_width(
    Team ~ px(35),
    everything() ~ px(55)
  ) %>%
  tab_footnote(
    footnote = "ICs on opponent or INCs on team",
    locations = cells_column_labels(
      columns = For
    )
  ) %>%
  tab_footnote(
    footnote = "INCs on opponent or ICs on team",
    locations = cells_column_labels(
      columns = Against
    )
  )