grab_bb.R

#' Obtain Exit Velocity and Batted Ball Distance Data
#'
#' @param id a "game_pk" id.
#' @examples
#' # Obtain data for a particular day
#' obs <- XML2R::XML2Obs("http://gd2.mlb.com/components/game/mlb/year_2015/month_07/day_08/miniscoreboard.xml")
#' gms <- obs[grepl("^games//game$", names(obs))]
#' ids <- XML2R::collapse_obs(gms)[, c("game_pk", "gameday_link")]
#' ids <- data.frame(ids, stringsAsFactors = FALSE)
#' ids$gameday_link <- paste0("gid_", ids$gameday_link)
#' bbs <- plyr::ldply(ids[, "game_pk"], grab_bb)
#' bbs <- plyr::join(bbs, ids)
#' 
#' # Now, add the exit velocity and distance to PITCHf/x data obtain by pitchRx
#' dat <- pitchRx::scrape(game.ids = ids$gameday_link)
#' dat$atbat <- plyr::join(dat$atbat, bbs, by = c("num", "gameday_link"))
#' 

grab_bb <- function(id) {
  feed <- sprintf("http://statsapi.mlb.com/api/v1/game/%s/feed/color", id)
  t <- jsonlite::fromJSON(feed, simplifyVector = FALSE)
  get_des <- function(x) {
    des <- x$data$description
    if (is.null(des)) "" else des
  }
  get_id <- function(x) {
    id <- x$id
    if (is.null(id)) "" else id
  }
  des <- unlist(lapply(t$items, get_des))
  # keep just the descriptions of play results
  # (this should hopefully match the number of at-bats)
  des <- des[unlist(lapply(t$items, get_id)) == "playResult"]
  
  idx <- grepl("[0-9]{2,3} mph", des)
  exit <- as.numeric(sub(".* ([0-9]{2,3}) mph.*", "\\1", des[idx]))
  
  idx2 <- grepl("[0-9]{1,3} feet", des)
  dis <- as.numeric(sub(".* ([0-9]{1,3}) feet.*", "\\1", des[idx2]))
  
  # start a matrix of missing values with one column for exit velocities,
  # one column for distance traveled (and one row for each play)
  m <- matrix(rep(NA, 2 * length(des)), ncol = 2)
  m[idx, 1] <- exit
  m[idx2, 2] <- dis
  df <- setNames(data.frame(m), c("exit", "distance"))
  N <- nrow(df)
  if (N > 0) {
    df$game_pk <- id
    df$num <- seq_len(N)
  }
  df
}

Could this be modified to get launch and/or spray angle as well?

Second question ...

I'm following along at https://baseballwithr.wordpress.com/2015/07/15/obtaining-exit-velocity-and-distance-of-batted-balls/. Thank you for these examples. I'm stuck on one part though. When I run:

bbs <- plyr::ldply(ids[, "game_pk"], grab_bb)

I get this error:

Error: lexical error: invalid char in json text.
                                       http://statsapi.mlb.com/api/v1/
                     (right here) ------^ 

This is with R 3.3.0 on OS X. I understand the intent of the error message but am unclear on what exactly the violation is and how to fix it. Do you have any advice on how to address this? Thank you!