Launch velocity and pitch location study

launchvelocity.R

# Load in useful packages

library(baseballr)
library(dplyr)
library(ggplot2)

# Read in the Statcast cata

b <- scrape_statcast_savant_batter_all("2017-08-04", 
                                       "2017-08-06")

# filter to only include pitches corresponding to batted balls

bip2 <- dplyr::filter(b, events %in% c("double", "double_play",
                               "field_error", "field_out",
                               "fielders_choice",
                               "fielders_choice_out",
                               "force_out",
                               "grounded_into_double_play",
                               "home_run",
                               "sac_fly", "single",
                               "triple"))


# define a ggplot2 theme and define the strike zone data frame

TH <- theme(plot.title = element_text(colour = "blue", size = 18, 
                      hjust = 0.5, vjust = 0.8, angle = 0))
topKzone <- 3.5
botKzone <- 1.6
inKzone <- -0.85
outKzone <- 0.85
kZone <- data.frame(
  x=c(inKzone, inKzone, outKzone, outKzone, inKzone),
  y=c(botKzone, topKzone, topKzone, botKzone, botKzone)
)

# convert the pitch location variables to numeric

bip2 <- dplyr::mutate(bip2,
               px = as.numeric(as.character(plate_x)),
               pz = as.numeric(as.character(plate_z)))

# initial scatterplot of pitch locations

ggplot(bip2, aes(px, pz)) + geom_point() +
  geom_path(aes(x, y), data=kZone, lwd=1, col="blue") +
  ggtitle("Pitch Locations of All Batted Balls") + TH

# same scatterplot where color corresponds to launch speed

bip2$Speed <- cut(bip2$launch_speed, 
                  c(0, 87.1, 96.4, 102.3, 120))
ggplot(bip2, aes(px, pz, color=Speed)) + 
  geom_point() +
  scale_color_brewer(palette="Set2") +
  geom_path(aes(x, y), data=kZone, lwd=1, col="blue") +
  ggtitle("Pitch Locations Colored by Launch Speed") + TH

# fit a generalized additive model using the gam() function

library(mgcv)
fit <- gam(launch_speed ~ s(px, pz), data=bip2)

# find predicted launch speeds over a grid of values

# the construct_grid() function sets up the grid of points

construct_grid <- function(xlo, xhi, ylo, yhi, npt){
  px0 <- seq(xlo, xhi, length.out=npt)
  pz0 <- seq(ylo, yhi, length.out=npt)
  PX <- outer(px0, rep(1, npt))
  PZ <- outer(rep(1, npt), pz0)
  data.frame(px=c(PX), pz=c(PZ))
}
dnew <- construct_grid(-1.5, 1.5, 1, 4, 50)

# find the predicted values

dnew$FIT <- predict(fit, dnew)

# construct a contour plot 

ggplot() + 
  geom_contour(data=dnew, 
          aes(x=px, y=pz, z=FIT),
          breaks=c(75, 80, 85, 90), color="red") +
  geom_path(aes(x, y), data=kZone, lwd=1, col="blue") +
  ggtitle("Contour Graph of Launch Speeds 
  Values at 90, 85, 80, 75") + TH