A while ago I listed some of my works in progress. Last week I unpaused one of them (again). Namely, making a database and interface for interacting with all of my running, cycling and hiking .gpx and .fit files. It feels like quite a big job. So I am going to endeavour to break it into small(ish), manageable(ish) chunks.

constraints

• it won't be on the internet^[1]
  • Local is fine
• it wont't have the ability to auto-magically sync^[2]
  • Manual will have to do is fine^[3]
• it won't neccessarily handle all of the data that my gps device captures
  • I'm ok not knowing what my heart rate was at 19:33:10 (BST) on the 22nd August 2019^[4]

functionality

• ingest my archive of several thousand .gpx & .fit files
• add new files as and when
• display an overview of a given ride
• search for tracks by date/time & by location
• display a heatmap sort of thing
• generate summary statistics and visualizations
• easily queryable

interface

A graphical interface is, I think, beyond my skillset. Instead I'm opting for a command-line interface (CLI). A text-based user interface (TUI) would be nice, but that too might be a stretch.

Aim low.

tools

• python because I know how to use it
  • with help from geopandas, shapely & movingpandas (maybe)
• duckdb because I want to learn how to use it
  • with it's SPATIAL extension
• uniplot because it looks like a good candidate for making nice plots in the terminal

The above will (& and some others, no doubt, at some point) will hande the functionality. And the below will help with the interface.

• typer because it seems like a good option for making a CLI; it handily makes --help documentation, and allows for fancy things like auto-complete
• poetry for bundling it all up into a package, whilst also managing dependencies

Many of these things are new to me.

current status

footnotes