Introduction
This short guide focuses on using espnscrapeR or the
nflverse/espnscrapeR-data repo to access QBR data.
Setup
If you have never installed the necessary R packages, go ahead and expand the collapsed section below, otherwise skip ahead to the “Load and Prep” stage.
Package Installation
You’ll need the following packages to get started. Note that as of
now, espnscrapeR is not on CRAN so you’ll need to install
it from GitHub as seen below.
install.packages(c("tidyverse", "gt", "remotes"), type = "binary")
remotes::install_github("espnscrapeR")Load and Prep
Go ahead and load the packages to get started.
library(espnscrapeR)
library(tidyverse)
#> ── Attaching packages ────────────────────────────────── tidyverse 1.3.2.9000 ──
#> ✔ ggplot2 3.3.6 ✔ dplyr 1.0.99.9000
#> ✔ tibble 3.1.8 ✔ stringr 1.4.1
#> ✔ tidyr 1.2.1 ✔ forcats 0.5.1
#> ✔ readr 2.1.3 ✔ lubridate 1.8.0
#> ✔ purrr 0.3.5
#> ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
#> ✖ dplyr::filter() masks stats::filter()
#> ✖ dplyr::lag() masks stats::lag()
library(gt)You can get the data directly from ESPN’s API.
# season level data (1x row per QB per season)
qbr_2020 <- get_nfl_qbr(2020, week = NA)
#> Scraping QBR totals for 2020!But it’ll be easier and recommended to just read in the data directly
with either nflreadr or just the raw URL.
nfl_qbr_season <- readr::read_csv("https://raw.githubusercontent.com/nflverse/espnscrapeR-data/master/data/qbr-nfl-season.csv")
nfl_qbr_season <- nflreadr::load_espn_qbr("nfl", seasons = 2006:2020)This is the QBR values for all QBs at the season level from 2006 to
now. The dplyr::glimpse() function can be used to quickly
see the type of the columns (IE numeric, character, etc) and the top few
values. You can think of it as a beefed up version of the
str() function.
nfl_qbr_season %>%
glimpse()
#> Rows: 1,102
#> Columns: 23
#> $ season <dbl> 2006, 2006, 2006, 2006, 2006, 2006, 2006, 2006, 2006, 20…
#> $ season_type <chr> "Regular", "Regular", "Regular", "Regular", "Regular", "…
#> $ game_week <chr> "Season Total", "Season Total", "Season Total", "Season …
#> $ team_abb <chr> "IND", "NE", "SD", "CIN", "NO", "BAL", "NYJ", "DAL", "PH…
#> $ player_id <chr> "1428", "2330", "5529", "4459", "2580", "733", "2149", "…
#> $ name_short <chr> "P. Manning", "T. Brady", "P. Rivers", "C. Palmer", "D. …
#> $ rank <dbl> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0,…
#> $ qbr_total <dbl> 86.4, 68.6, 67.4, 67.1, 66.7, 66.0, 64.2, 63.5, 62.1, 60…
#> $ pts_added <dbl> 85.5, 30.9, 28.2, 29.9, 36.7, 27.2, 20.8, 22.0, 17.2, 8.…
#> $ qb_plays <dbl> 624, 610, 542, 623, 631, 548, 587, 414, 380, 447, 717, 5…
#> $ epa_total <dbl> 108.8, 57.9, 53.0, 58.3, 64.2, 51.8, 47.1, 40.9, 34.9, 2…
#> $ pass <dbl> 96.0, 38.8, 43.1, 43.2, 61.0, 38.2, 27.9, 34.0, 17.1, 17…
#> $ run <dbl> 6.8, 4.3, -0.9, -0.3, -5.2, 8.2, 3.0, -2.9, 9.1, -0.3, 0…
#> $ exp_sack <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ penalty <dbl> 1.1, 2.8, 0.3, 2.5, 0.6, -0.6, 5.2, 0.5, -0.4, 2.4, 3.0,…
#> $ qbr_raw <dbl> 87.4, 67.2, 67.6, 66.4, 69.5, 66.9, 62.3, 67.9, 65.5, 56…
#> $ sack <dbl> -5.0, -12.0, -10.4, -12.9, -7.7, -6.0, -10.9, -9.2, -9.2…
#> $ name_first <chr> "Peyton", "Tom", "Philip", "Carson", "Drew", "Steve", "C…
#> $ name_last <chr> "Manning", "Brady", "Rivers", "Palmer", "Brees", "McNair…
#> $ name_display <chr> "Peyton Manning", "Tom Brady", "Philip Rivers", "Carson …
#> $ headshot_href <chr> "https://a.espncdn.com/i/headshots/nfl/players/full/1428…
#> $ team <chr> "Colts", "Patriots", "Chargers", "Bengals", "Saints", "R…
#> $ qualified <lgl> TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TR…Work with the data
Group By
We can group_by() the season and find the median QBR per
season.
nfl_qbr_season %>%
group_by(season) %>%
summarize(qbr_median = median(qbr_total), .groups = "drop")
#> # A tibble: 15 × 2
#> season qbr_median
#> <dbl> <dbl>
#> 1 2006 59.8
#> 2 2007 64.4
#> 3 2008 62.5
#> 4 2009 68.2
#> 5 2010 62.0
#> 6 2011 60.5
#> 7 2012 63.5
#> 8 2013 63.8
#> 9 2014 61.8
#> 10 2015 58.3
#> 11 2016 65.0
#> 12 2017 56
#> 13 2018 63.4
#> 14 2019 64.1
#> 15 2020 71.1We can also group_by() the season and find the max
n values per season.
top_16_per_yr <- nfl_qbr_season %>%
filter(qb_plays >= 100) %>%
select(season, team_abb, name_short, qbr_total) %>%
# group by season
group_by(season) %>%
# get top 16
slice_max(order_by = qbr_total, n = 16) %>%
# add the grouped median
mutate(qbr_median = median(qbr_total)) %>%
ungroup()
top_16_per_yr
#> # A tibble: 242 × 5
#> season team_abb name_short qbr_total qbr_median
#> <dbl> <chr> <chr> <dbl> <dbl>
#> 1 2006 IND P. Manning 86.4 66.4
#> 2 2006 NE T. Brady 83 66.4
#> 3 2006 IND P. Manning 71.9 66.4
#> 4 2006 NE T. Brady 68.6 66.4
#> 5 2006 SD P. Rivers 67.4 66.4
#> 6 2006 CIN C. Palmer 67.1 66.4
#> 7 2006 PHI J. Garcia 67.1 66.4
#> 8 2006 NO D. Brees 66.7 66.4
#> 9 2006 BAL S. McNair 66 66.4
#> 10 2006 TB T. Rattay 65.2 66.4
#> # … with 232 more rowsWe can then visualize this with a quick ggplot.
top_16_per_yr %>%
ggplot(aes(x = season, y = qbr_total, group = season)) +
geom_boxplot(alpha = 0.5) +
geom_jitter(width = 0.2, alpha = 0.5) +
geom_point(aes(y = qbr_median), color = "red", size = 3) +
theme_minimal()
Alternatively you can also find the median by quarterback.
nfl_qbr_season %>%
filter(qb_plays >= 100) %>%
group_by(name_short) %>%
summarize(
median = median(qbr_total),
years = range(season) %>% paste0(collapse = "-"),
active = if_else(max(season) == 2020, "Active", "Retired"),
.groups = "drop"
) %>%
arrange(desc(median))
#> # A tibble: 135 × 4
#> name_short median years active
#> <chr> <dbl> <chr> <chr>
#> 1 P. Mahomes 80.6 2018-2020 Active
#> 2 P. Manning 76.9 2006-2015 Retired
#> 3 T. Brady 73.9 2006-2020 Active
#> 4 L. Jackson 73.7 2018-2020 Active
#> 5 Q. Gray 73.6 2007-2007 Retired
#> 6 D. Prescott 71.9 2016-2020 Active
#> 7 T. Collins 71.1 2007-2007 Retired
#> 8 D. Brees 70.8 2006-2020 Active
#> 9 D. Watson 70.5 2017-2020 Active
#> 10 A. Rodgers 69.6 2008-2020 Active
#> # … with 125 more rows