R/similarityToolPitching.R

In Metlover/similarityTool: Using Fangraphs Data, Generates Similarity Scores for Batters

similarityToolPitching <- function(battedBallVals, playerName, numResults = 10, year=2017){
  playerName = paste(playerName, as.character(year), sep='')
  numResults = numResults + 1
  if(numResults > nrow(battedBallVals)) {numResults = nrow(battedBallVals) - 1}
  battedBallVals$Name = paste(battedBallVals$Name, as.character(battedBallVals$Season), sep='')
  names = battedBallVals[,"Name"] 
  ndx = which(names == playerName) 
  if("FNN" %in% rownames(installed.packages()) == FALSE) {install.packages("FNN")} 
  library(FNN)
  convertedVals<- apply(battedBallVals[,c('LD%','GB%','FB%','Pull%','Cent%','Oppo%','Soft%','Med%','Hard%','HR/FB','K/9','BB/9')],2, function(x){
    as.numeric(sub("%", "", x, fixed=TRUE))/100 }) 
  bbvNorm <- scale(convertedVals) 
  results = get.knnx(bbvNorm, as.data.frame(t(bbvNorm[ndx,])), k=numResults+1, algorithm="kd_tree") 
  resultingNames = names[results$nn.index[c(2:numResults)],] 
  endResult = data.frame ("Names" = resultingNames, "Scores" = results$nn.dist[c(2:numResults)], stringsAsFactors = FALSE)
  return(endResult) 
}