This repo contains the code I used to export my walking distance data from an iOS device and feed it to a custom-built system with a live dashboard and some nice visualizations.
I find the health dashboard on iOS to be annoyingly limited and I've been thinking about building something more useful for a while now. I've been eyeing Convex for some time as well, so this felt like a nice opportunity to take it for a spin.
The setup is fairly simple.
- An automated shortcut sends data to a Convex http endpoint
- Whenever it receives new data, Convex processes it and writes it to a table
- The table is hooked up to a React app that is automatically updated
First, export the data from the Health App of your iOS device. The export should look something like this:
├── export.xml
└── export_cda.xml
export.xml includes all the health data your device has collected up to this point. For my purposes, I am only interested in walking/running distance, but other metrics can easily be parsed by extending parse.py
To get the time series data (grouped by day), run:
python parse.py --path=<path-to-export.xml>
Install and start the web app:
cd app
npm install
npx convex dev # this will prompt you to set up convex
npm run dev
Import the data into convex:
npx convex import --table data <path-to-data.json>
To get live updates, you'll need an iOS shortcut that exports the data and sends it over to Convex. Here's a link to the one I'm using: https://www.icloud.com/shortcuts/b7be795b23f5455f8c955e976e052091 - very far from optimal, but it gets the job done and I didn't want to spent too much time on that part.
Note
As of May 2024, iOS shortcut automations can only be set up to run daily, meaning that if you need something to run every X minutes, you need to create multiple automations.
Note
This setup is best suited for a hobby personal project as it uses full table scans to fetch the data. Consider tweaking the schema or switching to a specialized time-series database if you need something more scalable.
Useful links: