Getting Over It

Visualising my Transalp bike ride

In summer 2020, I crossed the Alps with my road bike. I’ve recorded the whole ride and as a nice memory, I would like to visualise this ride.

Data

I collected the data using my GPS head unit and uploaded it to Strava. In this analysis it is available in its entirety as two .rds files.

To load, wrangle and visualise the data, the following libraries are used:

library(tarchetypes)
library(geotargets)
library(conflicted)
library(patchwork)
library(tidyterra)
library(tidyverse)
library(grateful)
library(elevatr)
library(ggrepel)
library(targets)
library(ggtext)
library(magick)
library(raster)
library(scales)
library(terra)
library(pins)
library(fs)
library(sf)

conflicts_prefer(dplyr::filter)
conflicts_prefer(dplyr::select)
conflicts_prefer(dplyr::lag)

theme_set(theme_minimal())

Define where exactly the data is located.

act_file <- "data/act.rds"
meas_file <- "data/meas.rds"

Read activity data from rds file.

• Define point of interests names
• Define if point of interest should be mapped to the start or the end of an activity
• Combine poi info for a string representation

act <- function(act_file) {
  read_rds(act_file) |>
    mutate(
      start_name = case_when(
        id == "3650448726" ~ "Albstadt",
        id == "3654245140" ~ "Winterthur",
        id == "3659045033" ~ "Fluelen",
        id == "3664650034" ~ "Andermatt",
        id == "3669729902" ~ "Andermatt",
        TRUE ~ NA_character_),
      end_name = case_when(
        id == "3650448726" ~ "Winterthur",
        id == "3654245140" ~ "Fluelen",
        id == "3659045033" ~ "Andermatt",
        id == "3664650034" ~ "Andermatt",
        id == "3669729902" ~ "Lugano",
        TRUE ~ NA_character_),
      poi_name = if_else(end_name == "Lugano", "Lugano", start_name),
      name = str_glue("{start_name} - {end_name}"),
      lat_poi = if_else(
        poi_name == "Lugano",
        map_dbl(end_latlng, 1),
        map_dbl(start_latlng, 1)),
      lng_poi = if_else(
        poi_name == "Lugano",
        map_dbl(end_latlng, 2),
        map_dbl(start_latlng, 2))) |>
    select(-distance)
}

df_act <- act(act_file)

Read measurement data. This data represents each point in time during my trip across the Alps. This includes among other things:

• Altitude
• Speed
• Longitude / Latitude
• Temperature
• Distance
• Cumulated Distance (Calculated)

meas <- function(meas_file) {
  df_meas_raw <- read_rds(meas_file)

  df_distance_max <- df_meas_raw |>
    group_by(id) |>
    summarise(distance_max = max(distance)) |>
    mutate(
      distance_max = lag(distance_max, default = 0),
      distance_max = cumsum(distance_max))

  df_meas_raw |>
    left_join(df_distance_max, by = join_by(id)) |>
    group_by(id) |>
    mutate(distance_cum = distance + distance_max) |>
    ungroup()
}

df_meas <- meas(meas_file)

Turn measurements into an sf (Pebesma and Bivand 2023) object:

act_meas_points <- function(df_meas, df_act) {
  df_meas |>
    st_as_sf(coords = c("lng", "lat"), crs = 4326) |>
    left_join(df_act, by = join_by(id))
}

sf_act_meas_points <- act_meas_points(df_meas, df_act)

The sf object represents the ride as point data. We combine it into one LINESTRING:

act_meas_lines <- function(sf_act_meas_points) {
  sf_act_meas_points |>
    group_by(id, name, lng_poi, lat_poi, poi_name) |>
    summarise(
      mean_speed = mean(velocity_smooth),
      mean_altitude = mean(altitude),
      mean_distance = mean(distance_cum),
      do_union = FALSE, .groups = "drop") |>
    st_cast("LINESTRING")
}

sf_act_meas_lines <- act_meas_lines(sf_act_meas_points)

With the help of the ‘elevatr’ package (Hollister et al. 2025), get the elevation data of the trip. Turn it into an ‘raster’ (Hijmans 2025) object for later plotting.

alpen_raster <- function(sf_lines) {
  get_elev_raster(sf_lines, z = 6, clip = "bbox", expand = 0.1) |>
    rast()
}

raster_alpen <- alpen_raster(sf_act_meas_lines)

Visualisation

Start with creating a map of the ride.

• Plot the sf object and add the raster data as a contour
• Map the id as color
• Add labels of the point of interests

vis_ride <- function(sf_act_meas_lines, raster_alpen) {
  ggplot(sf_act_meas_lines) +
    geom_spatraster_contour(data = raster_alpen, binwidth = 200) +
    geom_sf(aes(color = id)) +
    geom_label_repel(
      aes(x = lng_poi, y = lat_poi, label = poi_name),
      size = 2.5, fill = alpha(c("white"), 0.5)
    ) +
    labs(
      x = "Longitude", y = "Latitude"
    ) +
    theme(
      legend.position = "none", panel.grid.major = element_blank(),
      axis.text = element_blank(), axis.title = element_blank()
    )
}

If we were to create a plot with this function, the result would look like this:

The main challenge of the ride was the enormous elevation that was covered. Display this in an altitude plot:

• Filter for points in time where I was moving
• Create a line plot of the altitude
• Map the id as color
• Add an area beneath the altitude line
• Adjust scale and make sure limits start at 0 meters

vis_altitude <- function(sf_act_meas_points, sf_act_meas_lines) {
  sf_act_meas_points |>
    filter(moving) |>
    ggplot(aes(x = distance_cum, y = altitude, color = id, group = id, fill = id)) +
    geom_area(alpha = 0.6) +
    geom_line() +
    geom_label(
      data = sf_act_meas_lines,
      mapping = aes(x = mean_distance, y = mean_altitude, label = name),
      fill = alpha(c("white"), 0.5)
    ) +
    labs(x = "Distance [km]", y = "Height [m]") +
    theme(legend.position = "none") +
    expand_limits(y = 0) +
    scale_x_continuous(
      labels = label_number(scale = 0.001),
      breaks = breaks_width(50000)
    ) +
    scale_y_continuous(position = "right")
}

The plot would look like this:

Combine the two plots into one with the help of the patchwork (Pedersen 2025) package:

vis_transalp <- function(sf_act_meas_points, sf_act_meas_lines, raster_alpen) {
  gg_final <- vis_ride(sf_act_meas_lines, raster_alpen) +
    vis_altitude(sf_act_meas_points, sf_act_meas_lines) +
    plot_layout(widths = c(1, 3)) +
    plot_annotation(
      title = "Transalp 2020",
      subtitle = "Albstadt - Lugano"
    ) &
    theme(
      text = element_text(family = "Fira Code", size = 12)
    )

  ggsave(
    "transalp.png", gg_final,
    width = 40, height = 25, units = "cm", scale = 0.8
  )
}

gg_transalp <- vis_transalp(sf_act_meas_points, sf_act_meas_lines, 
     raster_alpen)

What a ride it has been! I’m always thinking back to this experience. It’s time to recreate such a ride in the near future.

Acknowledgments

This work was completed using R v. 4.4.1 (R Core Team 2024) and the following R packages: distill v. 1.6 (Dervieux et al. 2023), elevatr v. 0.99.1 (Hollister et al. 2025), fs v. 1.6.6 (Hester, Wickham, and Csárdi 2025), geotargets v. 0.3.1 (Tierney, Scott, and Brown 2024), ggrepel v. 0.9.6 (Slowikowski 2024), ggtext v. 0.1.2 (Wilke and Wiernik 2022), knitr v. 1.51 (Xie 2014, 2015, 2025), magick v. 2.9.0 (Ooms 2025), patchwork v. 1.3.2 (Pedersen 2025), pins v. 1.4.1 (Silge, Wickham, and Luraschi 2025), raster v. 3.6.32 (Hijmans 2025), rmarkdown v. 2.30 (Xie, Allaire, and Grolemund 2018; Xie, Dervieux, and Riederer 2020; Allaire et al. 2025), scales v. 1.4.0 (Wickham, Pedersen, and Seidel 2025), sf v. 1.0.23 (Pebesma 2018; Pebesma and Bivand 2023), shiny v. 1.12.1 (Chang et al. 2025), tarchetypes v. 0.13.2 (Landau 2021a), targets v. 1.11.4 (Landau 2021b), terra v. 1.8.93 (Hijmans 2026), tidyterra v. 0.7.2 (Hernangómez 2023), tidyverse v. 2.0.0 (Wickham et al. 2019).