timabe

timabe

statcast_cols <- cols(
  pitch_type = col_character(),
  game_date = col_date(format = ""),
  release_speed = col_double(),
  release_pos_x = col_double(),
  release_pos_z = col_double(),
  player_name = col_character(),
  batter = col_double(),
  pitcher = col_double(),
  events = col_character(),

timabe

load('busCats.Rdata') # load this data from github.com/timabe


catSums <- colSums(bus.cats) # get summary data, with just the category sums
catSums[order(-catSums)]->catSums # order the categories ahead of plotting them
plot(log(catSums))
# the plot shows a skewed set. Many of these will be useless in the hierarchical clustering
# as there are only a handful of observations of them. It's unlikely they will produce an 
# interesting cluster membership
# lets concentrate on the top 60 since there's a slight kink there

timabe

load('busCats.Rdata') # download this from https://dl.dropboxusercontent.com/u/6132890/busCats.Rdata

catSums <- colSums(bus.cats) # get summary data, with just the category sums
catSums[order(-catSums)]->catSums # order the categories ahead of plotting them
plot(log(catSums))
# the plot shows a skewed set. Many of these will be useless in the hierarchical clustering
# as there are only a handful of observations of them. It's unlikely they will produce an 
# interesting cluster membership
# lets concentrate on the top 60 since there's a slight kink there
# and for similarity purposes we only want businesses that have multiple categories

timabe

predictOverFullGames <- function(name) {
  player <- subset(pitch.test, pitcher_name == name)
  # genrate predictions as votes
  player$prediction  <- predict(rf, player, type = 'vote')
  # aggregate over his individual games
  playerAgg <- aggregate(. ~ V1, data = data.frame(cbind( player$url, player$prediction)), sum)
  # melt data frame to extract winner of majority vote
  melt.player <- melt(playerAgg, id = 'V1')
  melt.player$V1 <- as.character(melt.player$V1)
  melt.player <- ddply(melt.player, .(as.character(V1)), summarize, pitcher = variable[value == max(value)])

timabe

# Generate Training and Testing data from full pitch dataframe
train.vec2 <- sample(nrow(pitch), nrow(pitch) * 0.7)
pitch.train <- pitch[train.vec2, ]
pitch.test <- pitch[-train.vec2, ]

# Change pitcher name and pitch type from character variable to factor
pitch.train$pitcher_name <- as.factor(pitch.train$pitcher_name)
pitch.train$pitch_type <- as.factor(pitch.train$pitch_type)
pitch.test$pitcher_name <- as.factor(pitch.test$pitcher_name)
pitch.test$pitch_type <- as.factor(pitch.test$pitch_type)

timabe

# these variables aren't useful for analysis
badVars <- c('des', 'on_1b', 'on_2b', 'on_3b', 'type', 'type_confidence',
                'z0', 'zone', 'url', 'num', 'count', 'pitcher', 'batter_name', 
                'sv_id', 'nasty')

varsToOmit = c(badVars, 'pitcher_name', 'pitch_type')

# let's explore this data . . . use 3 pitchers and see if some dimentions of 
# their fastballs can differentiate them well. 

timabe

### HOT HAND 
###
### EXAMPLE: Anthony Morrow

# first source code from https://github.com/timabe/Hot-Hand/blob/master/hot_hand.R

streaks('Anthony Morrow')
### outputs it as data frame to streak_count
historicals('Anthony Morrow')
### outputs it as a list to historical.list