gt package - a (g)rammar of (t)ables
gt is an R package for generating formatted tables from dataframes in R with a Grammar of Tables.
If you want to go deeper than this basic guide, check out the gt site, which has lots of examples!
Raw data comes from: Pro Football Reference & Over the Cap
gt tablePer the package website, gt has the following component parts:
The parts (roughly from top to bottom) are:
- the Table Header (optional; with a title and possibly a subtitle)
- the Stub and the Stub Head (optional; contains row labels, optionally within row groups having row group labels and possibly summary labels when a summary is present)
- the Column Labels (contains column labels, optionally under spanner column labels)
- the Table Body (contains columns and rows of cells)
- the Table Footer (optional; possibly with footnotes and source notes)
As you can see it is fleshing out the idea of formatting or adding various parts of the table in a robust way.
Read in the Data
I’ve gone through collecting the data and have put into a non-tidy wide format for Salary Rank, playoff week and appearances, Total appearances, and finally salary from 2014-2019.
library(gt) # for static tables
library(tidyverse) # all the things
library(paletteer) # for all the palettes
playoff_salary <- read_csv("https://raw.githubusercontent.com/jthomasmock/radix_themockup/master/_posts/2020-05-13-qb-salaries-vs-playoff-appearances/playoff_salary.csv")
glimpse(playoff_salary)
Rows: 36
Columns: 7
$ player <chr> "Tom Brady", "Aaron Rodgers", "Russell Wilson", …
$ Wildcard <dbl> 1, 2, 4, 3, 2, 2, 2, 1, 2, 2, 2, 2, 1, 2, 0, 0, …
$ Division <dbl> 5, 4, 4, 3, 2, 2, 2, 2, 1, 2, 1, 1, 2, 0, 2, 2, …
$ Conference <dbl> 5, 3, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 2, 1, …
$ Superbowl <dbl> 4, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, …
$ Total <dbl> 15, 9, 10, 7, 4, 5, 6, 3, 3, 5, 3, 3, 5, 2, 5, 4…
$ salary <dbl> 100.065, 125.602, 91.114, 127.690, 93.700, 75.26…Basics of gt
A very basic gt table can be created as so:
playoff_salary %>%
head() %>%
gt()| player | Wildcard | Division | Conference | Superbowl | Total | salary |
|---|---|---|---|---|---|---|
| Tom Brady | 1 | 5 | 5 | 4 | 15 | 100.065 |
| Aaron Rodgers | 2 | 4 | 3 | 0 | 9 | 125.602 |
| Russell Wilson | 4 | 4 | 1 | 1 | 10 | 91.114 |
| Ben Roethlisberger | 3 | 3 | 1 | 0 | 7 | 127.690 |
| Alex Smith | 2 | 2 | 0 | 0 | 4 | 93.700 |
| Andrew Luck | 2 | 2 | 1 | 0 | 5 | 75.264 |
Immediately we have a basic table with minimal formatting.
The core parts we want to change are:
- Conditional color formatting for Total Appearances and Salary
- Change the fonts
- Add borders
Conditional Colors
We covered conditional colors in the previous post - check it out if you have further questions about building or using palettes.
A MAJOR value add that gt provides is rather than having to write our own palette parser it has a built in one via gt::data_color()! Just to refresh us on our palette, we’re using an adaptation of the viridis palette.
scales::show_col(c("#ffffff", "#f2fbd2", "#c9ecb4", "#93d3ab", "#35b0ab"))
For today’s example, we’ll use one from paletteer as well! Notice that printing paletteer::d() directly attaches the colors in the console, and to use it as a plain old character vector, we’ll need to wrap it in as.character() OR pipe it forward like palette %>% as.character().
paletteer::paletteer_d("ggsci::red_material", n = 6) %>%
as.character() %>%
scales::show_col()
I’d also like to introduce a function to generate colors from the scales package. scales::col_numeric() can generate continuous color scales based off a vector of colors. Notice that it operates differently than our other color function that is normalized between 0 and 1. This one takes ANY number and should be used with a vector of numbers. Notice the behavior with single colors generates the same color every time (nothing to scale against since we used domain = NULL).
red_color_generator <- scales::col_numeric(c("red", "white"), domain = NULL)
red_color_generator(50) %>% scales::show_col()