Visualizing Activities with activatr

In activatr: Utilities for Parsing and Plotting Activities

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

activatr (pronounced like the word "activator") is a library for parsing GPX files into a standard format, and then manipulating and visualizing those files.

Getting GPX Files

The process to get a GPX file varies depending on the service you use. In Garmin Connect, you can click the gear menu on an activity and click "Export to GPX". This package includes sample GPXs as examples.

Parsing GPX Files

Basic parsing of a GPX file is simple: we use the parse_gpx() function and pass it the name of the GPX file.

library(activatr)

# Get the running_example.gpx file included with this package.
filename <- system.file(
  "extdata",
  "running_example.gpx.gz",
  package = "activatr"
)

df <- parse_gpx(filename)

parse_gpx() returns an act_tbl, which has a column for latitude (lat), longitude (lon), elevation (ele, in meters), and time (time).

knitr::kable(head(df, 5))

activatr also overrides summary() to create a basic one-row tibble summarizing the activity.

summary(df)

knitr::kable(summary(df))

For more advanced parsing options, see vignette("parsing").

Analyzing GPX Files

Since this is just a tibble, we can analyze and plot it using usual techniques and libraries. activatr includes a few helpers, like mutate_with_speed(), speed_to_mile_pace() and pace_formatter() to make it easier to analyze pace using these libraries.

library(ggplot2)
library(dplyr)
df |>
  mutate_with_speed(lead = 10, lag = 10) |>
  mutate(pace = speed_to_mile_pace(speed)) |>
  filter(as.numeric(pace) < 1200) |>
  ggplot() +
  geom_line(aes(x = time, y = as.numeric(pace)), color = "blue") +
  scale_y_reverse(label = pace_formatter) +
  xlab("Time") +
  ylab("Pace (min/mile)")

For more details on those helpers, see vignette("pace").

Visualizing GPX Files

Once we have the data, it's useful to visualize it. While basic visualizations work as expected with a data frame:

library(ggplot2)
qplot(lon, lat, data = df)

It's more helpful to overlay this information on a map. To aid in that, get_ggmap_from_df() is a wrapper around ggmap::get_map() that returns a correctly sized and zoomed map, atop which we can visualize our track using ggmap::ggmap().

Let's see that on its own to start:

library(ggmap)
ggmap::ggmap(get_ggmap_from_df(df)) + theme_void()

# running_example_ggmap is the saved result of calling get_ggmap_from_df(df)
# We don't run that here because it requires an API key.
df_ggmap <- running_example_ggmap
ggmap::ggmap(running_example_ggmap) + theme_void()

We now have a map at the right size to visualize the run. Putting it all together, we can make a nice basic graphic of the run:

ggmap::ggmap(get_ggmap_from_df(df)) +
  theme_void() +
  geom_path(aes(x = lon, y = lat), linewidth = 1, data = df, color = "red")

ggmap::ggmap(running_example_ggmap) +
  theme_void() +
  geom_path(aes(x = lon, y = lat), linewidth = 1, data = df, color = "red")

activatr documentation built on Aug. 22, 2023, 9:12 a.m.