In petertea96/grandslamR: Ball Tracking Data in Professional Tennis

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

This is a brief tutorial on loading and plotting the courtvision tracking data. Please note that the grandslamR package is work in progress; if you have any suggestions or bugs to report, please let me know!

Package Installation

To install the package,

#install.packages('grandslamR')
#library(grandslamR)
library(ggplot2)
library(dplyr)
devtools::load_all()

Loading Data

There are two options from which you can load in data: - By a single match (using match_id) - By a single player (using player_id)

Loading Data by Match

Loading by a single match requires you to input a valid match_id. All valid match IDs can be found as follows:

knitr::kable(head(grandslamR::load_available_matches()))

Note that this data source includes a few doubles matches -- although the majority of available data are singles matches. Let's take a closer look at the 2019 Roland Garros final between Thiem and Nadal. To load in data from this match, we will need to know it's match_id.

thiem_nadal <- load_complete_match_data(match_id = 'atp_roland_garros_2019_SM001')
knitr::kable(head(thiem_nadal))

For a better understanding on what each column represents, please check the data dictionary.

Loading Data by Player Name

knitr::kable(head(load_complete_player_data(player_name = 'D.SHAPOVALOV')))

Plotting Data

The grandslamR package also supports plotting the spatio-temporal ball tracking data. For example, the following plot illustrates the ball bounce locations on a single rally point (1st set, 7th game, 12th point on 1st serve).

thiem_nadal %>%
  dplyr::filter(point_ID == '1_7_12_1') %>%
  dplyr::filter(position == 'bounce') %>%
ggplot2::ggplot(aes(x = x,
           y = y)) +
  gg_fulltenniscourt() +
  ggplot2::geom_point(alpha = 0.75, fill = 'green', shape = 21,
             size = 2.5) +
  ggplot2::geom_path(color = 'white', linetype = 2)