Zápisník Josefa Rouska

Python a React.js programátor, zaměřený na Shopify integrace

Visualizing Strava activites with Folium and Jupyter Notebook

Few days ago I participated in my first race ever! It was short triathlon called Slapský triatlon. My friend was talking me into it for almost two years. The swimming is 500m, cycling is 34km and running is 4.5km. I don’t swim or run that much but I was pretty sure I can handle this. I was pretty confident with cycling which turned out to be a mistake. I was happy with the result overall. I wasn’t last. I’m pretty satisfied with swimming and running, but I could push a bit harder on the bike. The result is that I have some interesting data I can play around.

Speed visualization of the cycling part. Ranging from deep blue for slowest to deep red for fastest.

When exploring the data I always use Jupyter Notebook and Folium. Folium lets you generate leaflet.js maps directly from Python.

Getting data from Strava is pretty simple. First, you need your Access Token. You can get them in My API Application section in Settings.

from stravalib.client import Client
client = Client(access_token = <STRAVA_ACCESS_TOKEN>)

types = ['time', 'distance', 'latlng', 'altitude', 'velocity_smooth', 'moving', 'grade_smooth']
streams = client.get_activity_streams(<ACTIVITY_ID>, types=types)

Then to get stream data use streams['altitude'].data. To make resulting map look a bit more smooth we need to group similar velocity data points. This does pretty good job.

locations = streams['latlng'].data
velocities = streams['velocity_smooth'].data

data = zip(locations, velocities)

groups = []
last_velocity = None
for location, velocity in data:
    if not last_velocity or abs(abs(last_velocity) - abs(velocity)) > 0.4:
        groups.append({'velocity': velocity, 'velocities': [], 'locations': []})
    groups[-1:][0]['locations'].append(location)
    groups[-1:][0]['velocities'].append(velocity)
    last_velocity = velocity

import statistics

for group in groups:
    group['velocity'] = statistics.median_high(group['velocities'])

To correctly colorize segments we need to normalize and map the values to colors. Luckily matplotlib does this for us with the exception of converting the color to hex. First, we will use Normalize to, well, normalize the value and then colormap value to color. The last step is converting rgb 0.0 - 1.0 to hex.

def convert_to_hex(rgba_color):
    red = str(hex(int(rgba_color[0]*255)))[2:].capitalize()
    green = str(hex(int(rgba_color[1]*255)))[2:].capitalize()
    blue = str(hex(int(rgba_color[2]*255)))[2:].capitalize()

    if blue=='0':
        blue = '00'
    if red=='0':
        red = '00'
    if green=='0':
        green='00'

    return '#'+ red + green + blue

import matplotlib.cm as cm
from matplotlib.colors import Normalize

cmap = cm.bwr
norm = Normalize(vmin=min(velocities), vmax=max(velocities))

def colorize(grade):
    return convert_to_hex(cmap(norm(grade)))

With grouped data and correct color the only thing we need to do is generate PolyLines for each group and display map.

import folium

m = folium.Map(location=[49.7947117,14.3916288], tiles='Stamen Toner', zoom_start=12.2)

for group in groups:
    folium.PolyLine(
        group['locations'],
        weight=10,
        color=colorize(group['velocity'])
    ).add_to(m)

m

You can see the entire notebook on nbviewer.jupyter.org. Here are my ride and run if you want to explore my activities on Strava.