Garmin Train

Python utilities for handling with Garmin workout data

Prequisites

• Python3
• Linux

Installation

1. Clone the repository
2. Install Python requirements
3. Give scripts executable permissions
4. Symlink (to update via git pull) or copy (for a one-time update) the Python scripts to /usr/local/bin

git clone https://github.com/addisonklinke/garmin-train.git
cd garmin-train
pip install -r requirements.txt
sudo python3 install.py --method symlink

Usage

1. Download activity data
   1. Login to your Garmin connect account
   2. From the left-hand panel, go to Activites > All Activites
   3. Click the name of the activity you are interested in
   4. Click the gear icon in the upper left and select "Export Original". This will download a .zip archive containing a .fit file with your heart rate and speed data (among other fields)
   5. (Optional) From the same menu, also select "Export GPX" to add elevation data
2. Convert the Garmin files into a CSV
3. Define a time range to analyze the CSV data for aerobic threshold (AeT) calculations

CSV Conversion

Starting in the folder with your downloaded activity files

user@linux:~/Downloads$ ls
5633299714.zip  5633299714.gpx

Use the activity2csv command to perform the conversion

• If you name your activity files with a common prefix, you can use the shell expansion * to pass them as a list
• -l/--locale is the only required flag and must represent a valid pytz.timezone string
• -g/--gaps is recommended to warn you of any large gaps in the timeseries data. Gaps longer than 1 minute may impact the accuracy of subsequent AeT calculations
• -n/--name adds an extra identifier to the resulting CSV. Without -n set, the default uses yyyymmdd convention to match the date of the activity

user@linux:~/Downloads$ activity2csv -g -l US/Mountain -n myActivity 5633299714.*

AeT Analysis

Use the rolling-aet command along with your newly converted CSV file

• Start and end times can use seconds-level precision if desired
• By default, tests use a 1-second rolling window with 30 minute halves. Either of these can be adjusted with the -f and -w flags, respectively
• Read more about the steps for conducting an AeT drift test in The Uphill Athelte's article
  • Sometimes outdoor courses have short downhill section(s) which should be ignored. This can be accomplished via one of the --max_speed/elev flags which will ignore times when the rolling (5-second) pace is greater than a threshold. You should pick the threshold by noting the live pace on your watch during these sections
• In an ideal test the pace is held constant, which means the drift in HR/Speed ratio will be identical to the drift in raw heart rate. However, if the average pace
  • Drifts higher: the HR/Speed drift will be LOWER than the raw HR drift. This indicates you were able to speed up while maintaining the same HR drift, so at a constant pace your HR drift would be lower
  • Drifts lower: the HR/Speed drift will be HIGHER than the raw HR drift. This indicates you could not keep up the same pace and had to slow down in order to attain the target HR drift

user@linux:~/Downloads$ rolling-aet -s HH:MM -e HH:MM yyyymmdd-myActivity.csv
+-----------------------------------------------------------------------------------------+
|                     Results from 59 windows: AeT 105.22 +/- 0.08 bpm                    |
+--------------+----------------+----------------+---------------------------+------------+
|    Method    | AeT Drift (%)  | Pace Drift (%) |         Pace @ AeT        | Successful |
+--------------+----------------+----------------+---------------------------+------------+
|     Raw      |  0.69 +/- 0.12 |       NA       |             NA            |    True    |
|   Hr/Speed   | -3.70 +/- 0.86 |  6.31 +/- 1.03 |     3.94 +/- 0.02 mph     |   False    |
| Hr/Elevation | -7.00 +/- 0.29 |  2.44 +/- 1.94 | 1231.84 +/- 12.81 ft/hour |   False    |
+--------------+----------------+----------------+---------------------------+------------+