Diagnostic Data Logger

As anyone who might have met me at PyCon UK over the years will probably have spotted (awful pun, fully intentional), I suffer from a facial skin condition called Roseacea.

Whilst there are several well-known triggers that cause people to have roseacea flare-ups, I’ve never managed to get to the bottom of what causes them for me.

So, inspired by a talk by Ian Ozsvald from a few years back (where he attempts to determine the cause of his wife’s sneezing), I thought I’d start capturing some data myself.

I put together an app that allows me to define the types of ‘observations’ that I’d like to record and then gives me a simple, quick interface that runs nicely on my smartphone.

The app took me around 30 minutes to write - and a large proportion of that was spent deciding which icons to use for each of the observation types!

Here’s a clone link for the end result:

https://anvil.works/build#clone:C5FBIJBOMHOFRB5F=XYRS6OZPRSZGAYQAZY4752MI

It’s dynamically driven by entries in its ‘observation_types’ data table. I’ve left the clone with the entries for my Roseacea investigation, just so you can see how it works, but you can change those however you wish to log something completely different. (You’ll also need to add a user record so you can log in).

Here’s hoping I can find some patterns in around a year or so!!

8 Likes

This is cool and shows why Anvil is so powerful. Icons always take the most time :joy:

For anyone that clones the app, you will need to add yourself as a user to the data tables to be able to run the app.

2 Likes

Updated with the option to capture images for any observation type - yep, I’m capturing the full gruesome nastiness in hi-res detail!

What would be interesting to build would be the addition of image recognition/analysis to match the observations over time. Thinking medically - it would be akin to drawing a line around an infection with a pen to see if it grows or shrinks :thinking: not sure if that has been done with this intention in mind but could be a valuable tool.

1 Like

Spot the spot, as it were!

image recognition aside, it would be relatively simple to add the captured photos to a canvas, allow the user to draw a bounding box and calculate a rough area from that (as long as the photos are sufficiently similar).