🚲 Bike Route Analyzer

What's the fastest way from A to B?

An app to analyze routes in gpx format to identify the most common starting and end points and analyze:

• duration of each routes
• speed
• "Crow Speed", i.e. the crow distance divided by time

... across:

• common start/end-point combinations (called "startendcluster" in the app)
• common routes for these combinations (called "clusters" in the app)
• temperature
• weekday
• season
• time of the day

The app is currently only tested with gpx files, recorded and downloaded from https://www.bikecitizens.net/de/, a wonderful app with easy recording functionality.

Screenshot

Installation

The python dependencies are shown in the file pyproject.toml. You can install and run with poetry:

poetry install
poetry run gunicorn app:app -b :8080

You can also install and run with pip:

pip install -r requirements.txt 
gunicorn app:app -b :8080

With python app.py, you can run the application in dash debug mode

Todos

Data

• test other sources than bike citizens
• data augmentation
• generate random test data
• include fitbit data

Analyzers

• tree models
• classification models
• SVC-algorithm, infer insights on features by probing predictions

GUI

• Accordion GUI for the workflow on the left panel
• Show single paths in a map when clicking on individual data
• Better coloring for Start and end
• Label Start/End clusters by getting addresses
• get average path for each cluster