MLS offseason turnover data, 2013 - 2015

analysis.R

data <- read.csv("turnover.csv")
summary(data)
str(data)


# Histograms

# Figure 1 - histogram of turnover by player
summary(data$Pctg.Players)
hist(
  data$Pctg.Players,
  xlim=c(0,1),
  col="grey",
  xlab="Percentage of players departing",
  main="Fig 1. Histogram of roster turnover, counting by player"
)
grid()

# Figure 2 - histogram of turnover by playing time
summary(data$Pctg.Minutes)
hist(
  data$Pctg.Minutes,
  xlim=c(0,1),
  col="grey",
  xlab="Percentage of Minutes departing",
  main="Fig 2. Histogram of roster turnover, counting by playing time"
)
grid()

# Scatterplot
plot(
  data$Pctg.Players, data$Pctg.Minutes,
  pch=19,
  las=1,
  col="black",
  xlim=c(0,1),
  ylim=c(0,1),
  main="Fig 3. Scatterplot of roster turnover",
  xlab="Percentage of players departing",
  ylab="Percentage of minutes departing"
)
grid()

# Tunover by team
plot(
  data$Team3,
  data$Pctg.Players,
  col="gray",
  las=1,
  ylim=c(0.2,0.7),
  main="Roster turnover by MLS team, measured by players",
  ylab="Percentage of players departing"
)
grid()

plot(
  data$Team3,
  data$Pctg.Minutes,
  col="gray",
  las=1,
  ylim=c(0.1,0.6),
  main="Fig 4. Roster turnover by MLS team, measured by playing time",
  ylab="Percentage of minutes departing"
)
grid()

# Comparing turnover to PPG earned
plot(
  data$PPG,
  data$Pctg.Minutes,
  col="black",
  las=1,
  pch=19,
  main="Fig 5. Roster turnover by minutes, compared to Points Per Game",
  xlab="Points per game",
  ylab="Roster turnover by minutes played"
)
grid()

# Modeling turnover based on team success
model <- lm(data$Pctg.Minutes ~ data$PPG)
summary(model)

# Individual team plots
par(mfrow=c(1,2))

plotTurnover("Chicago", colors.chi)
plotSuccess("Chicago", colors.chi)

plotTurnover("Columbus", colors.clb)
plotSuccess("Columbus", colors.clb)

plotTurnover("Colorado", colors.col)
plotSuccess("Colorado", colors.col)

plotTurnover("Dallas", colors.dal)
plotSuccess("Dallas", colors.dal)

plotTurnover("DC", colors.dc)
plotSuccess("DC", colors.dc)

plotTurnover("Houston", colors.hou)
plotSuccess("Houston", colors.hou)

plotTurnover("Kansas City", colors.kc)
plotSuccess("Kansas City", colors.kc)

plotTurnover("Los Angeles", colors.la)
plotSuccess("Los Angeles", colors.la)

plotTurnover("Montreal", colors.mtl)
plotSuccess("Montreal", colors.mtl)

plotTurnover("New England", colors.ne)
plotSuccess("New England", colors.ne)

plotTurnover("New York", colors.ny)
plotSuccess("New York", colors.ny)

# plotTurnover("New York City", colors.nyc)
# plotSuccess("New York City", colors.nyc)

# plotTurnover("Orlando", colors.orl)
# plotSuccess("Orlando", colors.orl)

plotTurnover("Philadelphia", colors.phi)
plotSuccess("Philadelphia", colors.phi)

plotTurnover("Portland", colors.por)
plotSuccess("Portland", colors.por)

plotTurnover("Real Salt Lake", colors.rsl)
plotSuccess("Real Salt Lake", colors.rsl)

plotTurnover("Seattle", colors.sea)
plotSuccess("Seattle", colors.sea)

plotTurnover("San Jose", colors.sj)
plotSuccess("San Jose", colors.sj)

plotTurnover("Toronto", colors.tor)
plotSuccess("Toronto", colors.tor)

plotTurnover("Vancouver", colors.van)
plotSuccess("Vancouver", colors.van)

custom_plots.R

plotTurnover <- function(team, color) {
  # Start new plot
  par(new=F)
  # Background dots are the entire league since 2013
  plot(
    data$Pctg.Players,
    data$Pctg.Minutes,
    pch=19,
    col="#dddddd",
    xlim=c(0.2,0.8),
    ylim=c(0.1,0.6),
    main="",
    xlab="% Players departing",
    ylab="% Minutes departing",
    las=1
  )
  # Next we draw the grid
  grid(col="#eeeeee")
  # Next we overlay a new dataset
  par(new=T)
  # Foreground dots are the team
  plot(
    data$Pctg.Players[data$Team==team],
    data$Pctg.Minutes[data$Team==team],
    pch=21,
    bg=color,
    col="black",
    xlim=c(0.2,0.8),
    ylim=c(0.1,0.6),
    main=paste(team, " Turnover"),
    xlab="",
    ylab="",
    las=1
  )
}

plotSuccess <- function(team, color) {
  # Start new plot
  par(new=F)
  # Background dots are the entire league since 2013
  plot(
    data$PPG,
    data$Pctg.Minutes,
    pch=19,
    col="#dddddd",
    xlim=c(0.5,1.9),
    ylim=c(0.1,0.6),
    main="",
    xlab="Points Per Game",
    ylab="% Minutes departing",
    las=1
  )
  # Next we draw the grid
  grid(col="#eeeeee")
  # Next we overlay a new dataset
  par(new=T)
  # Foreground dots are the team
  plot(
    data$PPG[data$Team==team],
    data$Pctg.Minutes[data$Team==team],
    pch=21,
    bg=color,
    col="black",
    xlim=c(0.5,1.9),
    ylim=c(0.1,0.6),
    main="Turnover by Success",
    xlab="",
    ylab="",
    las=1
  )
}

turnover.csv