NFL Tree and Cap Spending

Let’s explore the package ‘collapsableTree’ using NFL data. I foraged the internet to create a fairly simple csv file with columns Sport, Division, Region, Team, and Cap_Spend. The values in Cap_Spend are the team’s active (2020) active salary spends per OverTheCap.com.

Let’s get R set-up and then we can explore.

# Set CRAN Mirror
options(repos = c(CRAN = "http://cran.rstudio.com"))

# Set time zone
options(tz="America/New_York")

We’re going to use the collapsableTree library to create the visualization. Documentation is available here and here.

# Load data set 
library(readxl)
Football_Caps <- read_excel("~/Desktop/Football_Caps.xls")
head(Football_Caps)
# Load needed libraries
library(collapsibleTree)

The basic tree is fairly straight-forward to create.

# Click on nodes to expand
collapsibleTree(Football_Caps,
            hierarchy = c("Division","Region","Team"),
            width = 500)

(If the above image does not show on your screen, click HERE instead.)

Let’s also show the salary cap information that we added. We’ll need the dplyr package to do this. To get the cumulative spending for each branch of the hierarchy, we’ll use a sum function. Make sure you hover over each node and notice that the exact salary spending is listed.

library(dplyr)

Football_Caps %>%
  group_by(Division, Region, Team) %>%
  summarize(`Salary Spending` = sum(Cap_Spend)) %>%
collapsibleTreeSummary(
  hierarchy = c("Division","Region","Team"),
  root = "Football_Caps",
  width = 500,
  attribute = "Salary Spending")

What if we just wanted variation in the colors. I like blue, so we’ll have each node be a shade of blue.

divisionColors <- RColorBrewer::brewer.pal(length(unique(Football_Caps$Division)), "Blues")
# Regions will be a gradient that resets between divisions
regionColors <- Football_Caps %>%
  arrange(Division, Region) %>% 
  group_by(Division) %>%
  distinct(Region) %>%
  mutate(colors = colorspace::sequential_hcl(length(Region))[seq_along(Region)])
# Teams will also be a gradient that resets between divisions, but not Regions
teamColors <- Football_Caps %>%
  arrange(Division, Region) %>% 
  group_by(Division) %>%
  distinct(Team) %>%
  mutate(colors = colorspace::sequential_hcl(length(Team))[seq_along(Team)])

Football_Caps %>%
  arrange(Division, Region, Team) %>%
  collapsibleTree(
    hierarchy = c("Division", "Region", "Team"),
    root = "Football_Caps",
    width = 500,
    fill = c(divisionColors, regionColors$colors, teamColors$colors))

Let’s also examine the salary cap data. To first get a view of the spread, we’ll use a violin plot via the plotly package. Information on plotly is available here and here.

library(plotly)

Football_Caps %>%
  plot_ly(y = ~Cap_Spend, type = 'violin',
    box = list(visible = T),
    meanline = list(visible = T),
    x0 = 'Salary Spending') %>% 
  layout(yaxis = list(title = "", zeroline = F))

We see great dispersion in salary spending, ranging from $101M to 214M. The median is 154M USD. Let’s break this up between the AFC and NFC.

Football_Caps %>%
  plot_ly(x = ~Division, y = ~Cap_Spend, split = ~Division,
    type = 'violin',
    box = list(visible = T),
    meanline = list(visible = T)) %>% 
  layout(xaxis = list(title = "Division"),
         yaxis = list(title = "Salary Spending",zeroline = F))

The NFC tends to spend more on players than the AFC. Let’s look at individual teams.

plot_ly(Football_Caps, x = ~Team, y = ~Cap_Spend, type = 'bar') 
  %>% layout(title = "2020 Salary Spend in the NFL")

We see that the Jaguars spend the most while the Dolphins spend the least.

[THE END]

Home