Wednesday, March 23, 2016

Pebble Steps to MyFitnessPal

Back in January, I described how I sync my Pebble step count into my Fitbit account instead of using my Fitbit tracker. This week, I started using MyFitnessPal (MFP) to track calorie intake and I now need to incorporate these steps for calorie adjustment. The Rube-Goldberg machine gets bigger...

My first hope lay in the fact that MyFitnessPal app claims to integrate with iOS HealthKit. Since my steps are already in HealthKit, it seemed natural that I would just pull them from there into the MFP app. That proved to be a bust. Despite the allusion that one could hook up that data synchronization, both my testing and later confirmation with technical support indicated that this is not possible. In fact, I have not even been able to reliably sync calories out of  MFP into HealthKit; I get a couple of food records to go out. The integration appears to be half-baked.

My second approach was to use the fact that SyncSolver is transferring steps from HealthKit to Fitbit. MFP provides a direct pull from Fitbit. I crossed my fingers that this would work since steps pushed into Fitbit externally have limitations in terms of how they can be used. The winds blew favorably today and the steps from Fitbit do, in fact, eventually get into MFP. The key is "eventually." Due to syncing limitation, I don't actually see the steps in Fitbit until after about a 2 hour delay. Last night, I ate a meager dinner to fit within my calories, only to discover, as I was going to bed, that I could have had a dinner twice as big.($#*^&#$^ diet. However, until Fitbit figures out how to disable this, it does work. As a bonus, our Aria scale weights can be pulled straight from Fitbit into MFP via the same on-line connection.

Next trick, getting workouts into MFP ... Again, my attempt at HealthKit integration was a bust. However, iSmoothRun itself came to the rescue. I configured it to push directly to MFP on-line at the same time that it pushes to Dropbox and SportTracks. In principle, I may get trapped double counting my walks as "steps" and "workouts", but I can choose whether to push activities to MFP when I save. I am thinking I will not push walks and hikes but will push cycling activities that don't count as steps.

My Rube-Goldberg now looks like this:

As another note, I had been frustrated that automatic synchronization by SyncSolver were almost never happening and that I was having to manually push Sync Now in the app each day. Not a big deal with the corporate 10k-a-Day accounting ... much bigger deal now that my dinner depends on it! After contacting the authors, I found that SyncSolver is only able to pull from HealthKit when the phone is unlocked. Furthermore, it can only automatically do so when it receives a "Background App Refresh" notification from iOS. It turns out that the odds that a notification occurs when the phone is unlocked is pretty low as there is no coordination of these events. It gets worse ... The iOS has a smart feature to "learn" how often the target app needs to be refreshed. Since it discovers that SyncSolver is not actually succeeding when notified, it decides to reduce the frequency that it allows SyncSolver to try running.

Awesome. I can't imagine why this Apple HealthKit is not catching on like wildfire. The core problem would seem to be HIPAA. This law was supposed to make client health data more portable. Instead, it prevents use of our own health information. The law notwithstanding, this Apple behavior puzzles me. If I unlock the phone, I can manually perform the synchronization. In doing so, I have to trust that SyncSolver does not abuse its access. I also had to unlock the phone to install and configure SyncSolver in the first place. So, what damn difference does it make if the app does the actual synchronization when the phone is locked or unlocked. The app is not like a criminal that acts differently when someone is watching. I am not yet seeing the logic here.

Update Mar 27 - After deleting SyncSolver (beta) and reinstalling the newest version (v2.2), I am once again getting occasional automatic pushes. Perhaps Apple "unlearned" the desired refresh rate.

Sunday, March 13, 2016

Golden-cheeked Warblers Are Back

We took a walk on the Cactus Trail at Warbler Vista in the Balcones Canyonlands NWR this morning to look for our two Central Texas endangered species, the Golden-cheeked Warbler. Initially, I heard a half-dozen Black-and-white Warblers. Danielle saw one clearly but I never did. We final heard a couple of Golden-cheeked Warblers overhead in the budding Red Oak trees. Though I caught a quick naked eye glimpse, Danielle did not see it. Yay, they are back! In addition to these special birds, we found a few Titmice, Ruby-crowned Kinglets and Ladderback Woodpeckers. We also saw two Western Scrub-Jays which surprised me.

After lunch we went to try the Travis Audubon Baker Sanctuary. This is a members-only tract of land. It was a scenic walk with an nice babbling stream at the end of the loop. Perhaps it was because it was mid-afternoon but there was no bird life other than a Titmouse and a lone Orange-crowned Warbler. No pictures taken today; the camera and tripod were merely annoying ballast.

Update March 20 - Back to the Balcones Canyonlands NWR this morning. Once again a half-dozen Black-and-white Warblers along the trail. Still the best place I have been to see this species. There were more Golden-cheeked Warblers this time. I heard six or more and we saw three. Actually, all three were in the same tree which was somewhat surprising. Perhaps they have not staked out separate territories?

This time we managed to get some pictures of both the Black-and-white and Golden-cheeked Warblers. We also met Carlos Ross, one of the active birders from Austin.

Saturday, March 5, 2016

One Day - Two Rare Birds

It was crappy day ... ok, not in the way you are imagining. We started in late morning with a drive down to McKinney Falls State Park to the find the Red-Naped Sapsucker reported there this week. Based on some good descriptions posted on eBird, I had done some virtual scouting of the location on Google Earth. The bird was reported in "the only large Cottonwood tree" ... too bad I was not sure what a Cottonwood tree looks like without leaves! Eventually we figured it out by process of elimination and recorded its coordinates (30.1824526, -97.7254446). Here's a picture of the magical woodpecker tree, to the right of the trail:

We watched the tree for about 30min, finding two Downy Woodpeckers, a Red-bellied Woodpecker, and a Ladderback Woodpecker, all good omens. The spot we staked out was next to some fenced in equipment of unknown purpose. The longer we stood there the more evident it became that the purpose must be sewage.

Finally, we saw the sapsucker, a rarity in the Austin area and a new bird for us. This species is very similar to the Yellow-bellied Sapsucker, a more common winter resident here. This one is a female based on the white chin above the red throat, a pattern that does not appear on the yellow-bellied. The second picture shows the pale red patch at the nape that gives the bird its name. Sometimes, this spot is not present on female. The red-naped also does not have a solid black boarder around the chin and throat, another confirming field mark.

Red-Naped Sapsucker
Disappointingly, the sapsucker never came down low in the tree so our photography angles were awkward and interrupted by many unfortunately placed twigs.

On our way home we stopped by Hornsby Bend Bird Observatory. Yes, the waste water reclamation plant again. As Danielle asked, "what is it with birds and sewage." In Pond 2 we saw a large number of Northern Shoveler and Lesser Scaup. Mixed in were a handful of Pintails, Gadwalls, American Coots, and Ruddy Ducks.  Many of the Shovelers were feeding right at the northwest corner of Pond 2 and we got some shots as they flushed.

We then tried our luck at the mud flats at the west end of Pond 1W. We found a large number of Lesser Sandpipers huddled together on small patches of drier ground. There were quite a few Killdeer as well, more than I have seen before. We also saw a half-dozen Wilson's Snipes rapidly probing into the mud. Most were in the tall weeds surrounding the flats, this one was out in the open. Up until now, it had been exciting for us to find a single snipe somewhere!

While watching the mud flat, we encountered the bird shown below. At first we thought it was some sort of sandpiper until we saw how small the bill was. We concluded it was one of the taller-standing plovers of genus Pluvialis. The other plovers, of genus Charadrius, including the Killdeer look more squat and compact.

American Golden-Plover
We debated for some time as to which species it was since, in non-breeding plumage, they look similar. The two species with any likelihood of being found in Texas were the Black-bellied Plover and the American Golden-Plover. Both of these came up in eBird as rarities. We settled on the American Golden-Plover based mostly on the dark ear patch and the strongly contrasting cap. The shorter, thinner bill is also indicative, but I had nothing to compare to. The lack of black armpits would have been diagnostic but we did not catch it when it flew away. Just to be sure, we eliminated the possibility of a very rare Pacific Golden Plover based the shorter tertials and longer primaries relative to the tail. This is a new species for both of us and wrapped up our wonderfully crappy day.

Update Mar 06 - Our identification of the plover was confirmed on eBird and a several other birders have seen it today. Interestingly, none of those people had their sightings marked as rare, puzzling. I tinkered around with eBird and found that a plover seen on Saturday would have been marked as rare but not one on Sunday. The rarity is seasonal and presumably delineated in the database by week. Sunday would have been the start of the first week in March, when plover sightings historically begin. Looks like we scored a rare bird on a mere technicality, awesome.

Sync Solver Back on Track

I was able to TestFlight a beta version of the new Sync Solver Health to Fitbit app for iOS. Previously, I had posted about a problem with the correctness of daily totals that resulted from hourly totals being accumulated in the wrong slot. This upcoming version has fixed the problem. Now the hourly and daily totals in iOS Health (generated from my Pebble) exactly match those pushed into Fitbit.

Here is a screen shot of my Pebble Time watch this morning showing that yesterday, I had 17867 steps.

Next is a spreadsheet showing my steps from over the last 40 hours as exported from iOS Health using QS Access. Highlighted are the records from Friday. The cell highlighted in orange is the sum of these and matches the 17867 steps.

For comparison, here is a screen capture of my Fitbit account with all of the Friday, March 4 records highlighted as well as the daily total computed by Fitbit. Not only is the daily total 17867, all of the individual hourly values match and are in the correct slots.

Finally, the all important number reported to my employer through the 10k-a-Day site is also 17867.

I have not been able to test the similar problem that occurred during the initial push of historical values. Perhaps others might report their success.