fastball count (2013)

fastballcount.R

library(pitchRx)
library(ggplot2)
library(data.table)
library(dplyr)

# データ取得
# data = scrapeFX("2013-01-01", "2013-10-01")
# 取得済みデータの利用
# data <- fread("2013.csv")

# 公式戦のデータを利用
data <- subset(data, sv_id != "NA")
# 直球系ボールフラグ変数を追加
data$F_flag <- ifelse(data$pitch_type %in% c("FF", "FS"), 1, 0)
#投手名とカウントでデータを分割
data_count <- data %.% group_by(pitcher_name, count) %.%
  dplyr::summarise(pitches = length(pitch_type), fs = sum(F_flag))
# 直球率を計算
data_count$ratio <- round(with(data_count, fs / pitches), digit = 3)
# setkey
setkey(data_count, pitcher_name)
# subsetするよりも速いらしい
pitch_data <- data_count[J("Hiroki Kuroda"),]
pitch_data

# plot
gg <- ggplot(data = pitch_data) + geom_bar(aes(x=count, y = ratio, fill = "red"), alpha = 0.8, stat = "identity")
gg <- gg + theme(legend.position = "none") + ggtitle("Hiroki Kuroda")
gg <- gg + theme(plot.title = element_text(size = 20, face ="bold"))
gg