Strava Dashboard

A personal project to explore my Strava data

Motivation

There’s nothing better than having fun with data that’s close to your heart. For me, it’s my Strava data (yes, I admit that sounds a bit silly). For my sanity and to decompress after work I like to spend my free time for hicking, running or cycling. It’s also allowed me to learn more about APIs, graph creation…

The project is still a work in progress, but the first version already includes total mileage for hiking, biking, and running, as well as a heat map of all my activities.

There are many more visualizations and features I’d love to add, but finding the time has been a challenge. For now, I manually retrieve new activities using the Strava API via the {rstrava} package, but I plan to automate this in the next version. Stay tuned!

Code written by me, Corentin Maslard and inspired by various websites:

• https://www.dancullen.me/articles/creating-a-heatmap-in-r-with-google-polylines
• https://padpadpadpad.github.io/post/animate-your-strava-activities-using-rstrava-and-gganimate/
• https://medium.com/@annthurium/getting-started-with-the-strava-api-a-tutorial-f3909496cd2d
• https://www.r-bloggers.com/where-do-you-run-to-map-your-strava-activities-on-static-and-leaflet-maps/
• http://www.databrew.cc/posts/strava.html
• https://github.com/fawda123/rStrava

The rStrava package allows easy access to your Strava data using the Strava API. Combined with leaflet I can easily create a heatmap of all of my Strava activities. One interesting challenge does arise when doing this. Strava stores the coordinates of the activity as a Google Polyline, so I will use the googleway package to decode the data in to latitude and longitude coordinates.

Data importation

Then I will enter my personal Strava API information which can be found here after you apply for API access.

## Necessary info from Strava api https://www.strava.com/settings/api 
app_name       <- 'website' # chosen by user
app_client_id  <- '133545' # an integer, assigned by Strava
app_secret     <- 'e606d239de2a102bf125357c62641af7b2c39e04' # an alphanumeric secret, assigned by Strava

stoken <- httr::config(token = strava_oauth(app_name, app_client_id, app_secret, app_scope="activity:read_all"))

Next, I load in my activity list and compile my activities into a data frame using the rStrava library.

## Load my activities and compile activities with rStrava library
my_data  <- get_activity_list(stoken)
act_data <- compile_activities(my_data)
save(my_data, act_data,stoken, file = "perso_projects/media/strava_report/data_strava.RData")

The activity data contains 56 variables. We only need to keep two, the map data called ‘map.summary_polyline’ and the ‘upload_id’.

load(here::here("perso_projects/media/strava_report/data_strava.RData"))
## Keep only activity id and map line
keeps   <- c('map.summary_polyline', 'upload_id')
my_acts <- dplyr::select(act_data, match(keeps, names(act_data)))

Now that the data is prepared I want to create a blank map of the area I want to include in my map. I create longitude and latitude bounds to create the map area I want. This will take some trial and error to get the exact bounds that you want.

## Create blank map bounded by given lon and lat
# Dijon
lons.range <- c(5, 5.1) #(x)
lats.range <- c(47.2, 47.4) #(y)
## tile options CartoDB.Positron , CartoDB.DarkMatter , Stamen.Toner  
map <- leaflet(options = leafletOptions(zoomControl = TRUE)) %>%
  addProviderTiles('CartoDB.Positron',
                   options = providerTileOptions(noWrap = T, minZoom=0, maxZoom=50)) %>%
  fitBounds(lng1 = min(lons.range), lat1 = max(lats.range), lng2 <- max(lons.range), lat2 = min(lats.range))
#map   #save using mapshot(map, file="sbmap.png")

Next I loop through all my activities and convert them from Google Polylines to a dataframe of longitudes and latitudes each loop through will add one activity to the map.

Below is a heat map of all my recorded activities, created using leaflet.

# my_acts<-my_acts %>% 
#   drop_na() %>% 
#   dplyr::filter(map.summary_polyline!="")
# loop <- unique(my_acts$upload_id)
# for (i in loop) {
#   activity <- dplyr::filter(my_acts, upload_id == i)
#   coords   <- googleway::decode_pl(activity$map.summary_polyline)
#   map      <- addPolylines(map, lng = coords$lon, lat = coords$lat,
#                       color = 'red', opacity = 1/4, weight = 2)
# }
# 
# map

Change color depending of the different activity

my_acts_select <- get_activity_list(stoken, after = as.Date('2021-01-01'))
act_data <- compile_activities(my_acts_select) %>% 
  dplyr::filter(map.summary_polyline!="") %>% 
  mutate(type2 = ifelse(type %in% c("Ride", "Run", "Hike"), type, "Other"))

loop <- unique(act_data$upload_id)

for (i in loop) {
  activity <- dplyr::filter(act_data, upload_id == i)
  coords   <- googleway::decode_pl(activity$map.summary_polyline)
  
  color <- switch(activity$type2,
                  "Ride" = '#007FE0',
                  "Run"  = '#fc4c02',
                  "Hike" = '#08AF93',
                  "Other" = '#F15BB5')  # Default color
  
  map <- addPolylines(map, lng = coords$lon, lat = coords$lat,
                      color = color, opacity = 1/4, weight = 2,
                      group = activity$type2)  # Ajouter la polyligne à un groupe
}

# Ajouter les contrôles JavaScript pour le filtrage
map <- map %>% addLayersControl(
  overlayGroups = unique(act_data$type2),
  options = layersControlOptions(collapsed = FALSE)
)
map

#athl_fun(133545, trace = FALSE)
my_acts_select <- get_activity_list(stoken, after = as.Date('2021-01-01'))
act_data <- compile_activities(my_acts_select)
act_data<- act_data %>%
  dplyr::filter(type %in% c("Hike","Ride","Run","Workout")) %>% 
  dplyr::mutate(
    cumulative_distance = cumsum(distance),
    cumulative_duration = cumsum(moving_time),
    Year = format(as.Date(start_date), "%Y"),
    month = format(as.Date(start_date), "%m-%d")
  ) %>% 
  dplyr::group_by(Year, type) %>%
  dplyr::mutate(cumulative_duration_years = cumsum(moving_time),
                cumulative_distance_years = cumsum(distance)) %>%
  dplyr::ungroup() %>% 
  mutate(Hours = cumulative_duration_years/3600)

#test=act_data %>% dplyr::select(cumulative_duration_years,Hours,year,month)

Years <- unique(act_data$Year)

n_years <- length(Years)
my_gradient <- colorRampPalette(colors = c("#230007","#7F9CC7","#FFBA08", "#fc4c02"))(n_years)

p1<-act_data %>% 
  ggplot(aes(x= as.Date(month,"%m-%d"),y=Hours, col=Year, group = Year))+
  geom_line()+
  facet_wrap(type~., scales = "free_y")+
  ylab("Cumulative duration (h)")+
  xlab("Month")+
  scale_x_date(date_labels = "%m", date_breaks = "1 month") +  # Format pour les labels des mois
  theme_minimal()+
  scale_color_manual(values = my_gradient)

p2<-act_data %>% 
  ggplot(aes(x= as.Date(month,"%m-%d"),y=cumulative_distance_years, col=Year, group = Year))+
  geom_line()+
  facet_wrap(type~., scales = "free_y")+
  ylab("Distance (km)")+
  xlab("Month")+
  scale_x_date(date_labels = "%m", date_breaks = "1 month") +  # Format pour les labels des mois
  theme_minimal()+
  scale_color_manual(values = my_gradient)
#p1
ggplotly(p1)

#ggplotly(p2)