statcast_impute_derivation.R

library(baseballr)
library(dplyr)

obtain_data = function(type="postgres", 
                       start_date="2017-03-29", 
                       end_date="2017-10-03", infile=NULL) {
  
  if (type=="postgres") {
    # postgres db connection here
    # e.g.
    # library(RPostgres)
    # library(DBI)
    # conn <- dbConnect(RPostgres::Postgres(), 
    #  password=SOMEPASS, user=SOME_USER, port=SOME_PORT_PROBABLY_5432, dbname=SOME_NAME)
    # df1 = dbGetQuery(conn, "select * from SOME_TABLE_NAME where SOME_SELECTION_CRITERIA")
  } else if (type=="rds") {
    readRDS(infile)
  } else if (type=="csv") {
    read.csv(infile, stringsAsFactors = FALSE)
  } else if (type=="scrape") {
    date_seq = seq(as.Date(start_date), as.Date(end_date), by=1)
    dplyr::bind_rows(lapply(date_seq), function(d) {
      baseballr::scrape_statcast_savant(as.character(d), as.character(d))
    })
  }

} 

statcast_impute_derive = function(statcast_df, inverse_precision=10000) {
  # statcast_df must have columns launch_angle, launch_speed, bb_type, events
  
  aa = statcast_df %>% 
  filter(!is.na(launch_speed)) %>% 
  mutate(ila=round(launch_angle*inverse_precision), 
         ils=round(launch_speed*inverse_precision))
  
  la_ls_count = aa %>% group_by(ila, ils, bb_type, events) %>% summarise(n=n())
  
  # use n >= 5 here? some other number? 99.X percentile? this is why it's a heuristic
  la_ls_filtered = la_ls_count %>% select(ila, ils, bb_type, events) %>% filter(n>=5)
  
  la_ls_filtered %>% write.csv("CSV_FILE_TO_LOAD_LATER.csv", row.names = FALSE)
  
}