Use Exercise Markers in Prediction #13
Comments
|
This does only apply to x23 and up, but this is still a great feature! You'll need to open a similar feature in https://github.com/loudnate/mmhistorytools so we don't strip the exercise marker. We might also need to make sure https://github.com/bewest/decoding-carelink is parsing them first. |
|
Good ideas. Are there any samples of data with the desired markers in them with known values?
|
|
So here is the hexdump: CSV is at: Of note, the exercise marker is in between two meal markers, which do get decoded correctly. |
|
Nice. :) |
|
So, the marker has not numeric value or anything? What do you intend for this to mean, what does the system intend for this to mean, is there any difference? Is it similar to "eating soon mode" in terms of "there's exercise happening... " "already", "now," "soon." for "$duration" minutes? Of intensity $itense? Can we get this from fitbit or similar? Looks like, if I'm understanding correctly that this is a simple binary marker of "some exercise may or may not be happening."? Any proposed json export for this? Thanks again, great job, the hexdump is perfect. |
|
My use case is along the lines of "I am about to exercise," with each marker counting for a "unit" of exercise. The system would see it, and just like it superimposes carbs and insulin activity into the prediction, it would add it in. There would need to be a user configuration somewhere that specifies the BG equivalency of an "exercise unit" and the duration over which it is applied; for me, I'd probably use something like 20 mg/dL over 15 minutes. If I was going to do intense exercise, I would enter two markers in, one right after the other. If I wanted to do a long exercise, I would have to re-enter this every 15 minutes. That last part isn't ideal, so I'm happy to rethink how to better handle it. You're right that it is simply binary at the pump level, which is unfortunate—a duration entry would be great. Two different scenarios occurred that made me think this would be a good idea: one was biking home from a bar, where my BG was good but I knew it would drop. If I had entered an exercise marker, the system would have started to adjust earlier instead of waiting until I dropped. The other scenario was a hike that went most of the day. I kept fighting the system, essentially by eating without telling it I was eating. In that case, some way to indicate that I would be doing this for a longer duration would be better. FitBit data may work, but it would be lagging. |
|
opcode = 0x41 |
Much like meal markers are used in combination with the bolus wizard entries for COB analysis, it would be great to use the exercise marker to adjust the prediction. There are many ways to do this, but one is to predict a drop of x mg/dL over y min for each marker that is placed. Perhaps set it at a 1/2 U ISF equivalent or something (all could be user-configurable). When I'm about to bike or something, I can simply add a marker or two on the pump, which will shift the prediction and adjust accordingly. If I'm exercising longer, I can add more markers.
Might only apply to x23 series, not sure.
The text was updated successfully, but these errors were encountered: