R code for "cleaning" Statcast data

clean_statcast.R

# read in SC data from  2019 seasons
# through games of June 23, 2019

library(tidyverse)
library(CalledStrike)

sc2019 <- read_csv("~/Dropbox/2016 WORK/BLOG Baseball R/OTHER/StatcastData/statcast2019.csv")
sc2019_ip <- filter(sc2019, type == "X")

# focus on Statcast 2019 in-play data
# look for weird values of launch speed and launch angle

# first graph the distinct values of launch speed

sc2019_ip %>%
  group_by(launch_speed) %>%
  summarize(N = n()) %>%
  ggplot(aes(launch_speed, N)) +
  geom_col() +
  increasefont() +
  ggtitle("Distribution of 2019 Launch Speeds") +
  centertitle()

# graph of the distinct values of launch angles

sc2019_ip %>%
  group_by(launch_angle) %>%
  summarize(N = n()) %>%
  ggplot(aes(launch_angle, N)) +
  geom_col() +
  increasefont() +
  ggtitle("Distribution of 2019 Launch Angles") +
  centertitle()

# pick out the extreme pairs of (ls, la)
# finds the pairs where the frequency exceeds n_min

find_extreme <- function(d, n_min = 80){
  d %>%
    group_by(launch_speed, launch_angle) %>%
    summarize(N = n()) -> S
  S %>% filter(N >= n_min)
}

# decide on using a frequency of 100 as a cutoff

o19 <- find_extreme(sc2019_ip, n_min = 100) %>% 
  arrange(desc(N))

# remove these pairs from Statcast data
# must be a more elegant way of doing this

sc2019_ip %>% mutate(Row = row_number()) -> sc2019_ip
find_rows_one_pair <- function(j){
   sc2019_ip %>% 
    filter(launch_speed == o19$launch_speed[j],
           launch_angle == o19$launch_angle[j]) %>% 
    select(Row) %>% pull()
}

row_numbers <- NULL
for(j in 1:nrow(o19)){
  row_numbers <- c(row_numbers, find_rows_one_pair(j))
}

sc2019_ip %>%
  filter(!row_number() %in% row_numbers) ->
  sc2019_clean

# graphs of cleaned data

sc2019_clean %>%
  group_by(launch_speed) %>%
  summarize(N = n()) %>%
  ggplot(aes(launch_speed, N)) +
  geom_col() +
  increasefont() +
  ggtitle("Distribution of 2019 Launch Speeds") +
  centertitle()

# graph of launch angles from cleaned data

sc2019_clean %>%
  group_by(launch_angle) %>%
  summarize(N = n()) %>%
  ggplot(aes(launch_angle, N)) +
  geom_col() +
  increasefont() +
  ggtitle("Distribution of 2019 Launch Angles") +
  centertitle()