I thought at first to post this in the Dexcom group, but I'd like to get input from the general community.
What I'm doing is nothing new. Others have been working on this in various forms for years. In fact, it is leveraging their open-source decoding of the Dexcom USB protocol that is my starting point.
I have a Samsung S4 Android smart phone, and a Samsung Gear smartwatch. The latter runs Android too, has been hacked, so I'm running a full-blown Android Jellybean release on the watch.
I'm starting a personal project to develop an application and user interface for presenting Dexcom data on my phone, and my watch. Obviously the watch interface will be simpler, as there is less screen real-estate to work with.
My purpose in this posting is to solicit ideas for what the app can do with the data. I'll be implementing the same functionality as the receiver (display last reading, trend graph, trend arrow, settable limits, alarms, etc.).
The additional features/functionality that I'm already planning:
- Extended history -- the reciever has (I believe) 90 days of history stored. App will allow full review of this history, rather than just the last 24 hours.
- Moving average(s). MA trends
- Same-period comparison -- i.e. overlay the history for the period from 4AM-8AM each day; do some statistical comparisons (normalize, calculated correlation, etc.)
- Variable alarm limits through 24-hour period (i.e. set your low limit lower while your awake, bump it up at night while asleep, say 65/75 for example; or the "I need to sleep" profile with high limit 140 during the day, 160 while sleeping)
- (phase 2) Cloud-based storage/retrieval/notification
That very last one I may not bother with at all if I can integrate with similar efforts already under way. For the less tech-savvy, what the Cloud aspect means is that a parent can monitor their child's BG on their smartphone wherever they are (and call school authorities if necessary). Or, if an adult gets in a bad situation, like with a very serious hypo that they can't manage themselves, a loved-one can get some sort of alarm on their phone about it, and get them help.
I've already got the basics working thanks to work already done by John Costik (see his site, hackingtype1.com) and posted to GitHub. I expect this to be a several months-long project to get to the first "beta" release, as I'm doing it on my spare time.
However, I will, like John, make all my code available as open-source, so anyone that wants to make use of it is welcome to.