fitness-analysis

Analysis tools for fitness activity data from Strava and MyNetDiary.

Provides functions for loading and processing MyNetDiary and Strava fitness data into Pandas DataFrames that are convenient for analysis and plotting.

The module assumes Strava data are an unzipped directory of a downloaded user account archive from www.strava.com. For MyNetDiary, the module assumes a directory of separate MyNetDiary_Year_XXXX.xls user data files downloaded from www.mynetdiary.com.

Installation

pip install git+https://github.com/tabishm52/fitness_analysis.git

This pulls in activity-parser automatically as a declared dependency.

Usage

import fitness_analysis as fa

weight, calories = fa.load_mnd_data("path/to/MyNetDiary/files/", ...)
activities, weekly_sums = fa.load_strava_activities("path/to/Strava/archive/", ...)
power_curves = fa.load_power_curves("path/to/Strava/archive/", ...)

Other capabilities:

• load_commute_activities — summary metrics (distance, time, elevation) for recurring bike-commute activities. Distinct from load_strava_activities because it supports automatic splitting of round-trip commutes recorded as one activity.
• load_activity_records / load_activity_coords — load parsed FIT/TCX/GPX records or trimmed lat/lon data for a set of activity files.
• cluster_routes — cluster bicycle activities by GPS route similarity (Fréchet distance) or activity name.
• geocode_positions / seed_geocode_cache — reverse/forward geocode GPS positions into addresses.
• piecewise_fit_fixed / piecewise_fit_auto — piecewise linear regression on a time series, with automatic breakpoint/segment-count selection. piecewise_fit_cached wraps these with disk caching.

Most loaders cache their results to disk (Parquet or SQLite) and are paired with an invalidate_*_cache function to force a refresh.