I am quite curious about how spark comes up with the exercise calories. I know the exercise minutes are for the minutes fitbit classes as either fairly active or very active. I assumed the exercise calories are just the total fitbit calorie burn for those minutes. But after looking at it in detail, I see it isn't the case.
I am using yesterday as an example as I did a fair amount of walking, but did not log any activity on the fitbit site. Here are the sources I am comparing:
*The fitbit dashboard--calorie burn, and the activity dashboard that shows the pie chart for the minutes in the four activity levels.
*The dashboard on my Digifit account. Digiffit is a heart rate monitor app that can link with fitbit, one of the views on the Digifit dashboard is the calorie burn from fitbit for time spend sleeping, sedentary, lightly active, fairly active and very active. These all add up to equal my fitbit burn that day so I think it is accurately pulling the data from fitbit.
*On Spark, the exercise log (exercise minutes and calories) and the "Calories Burned Differential Report" on the Spark phone app.
The minutes work out just as I expected. Spark counted 182 minutes, and that equals my fitbit fairly active + very active minutes. So far so good.
Now Spark credits me with 819 calories burned that day. When I add up the fairly active and very active calories burned (from Digifit, fitbit doesn't break this down), that results in 784 calories burned. So it must not be based on that.
I also noticed the calorie burn Spark assumes for me is higher than my total calorie burn on fitbit. According to the Calorie differential report in the Spark app, Spark estimates I burned 2447 calories. My fitbit calorie burn estimate for the same day was 2058. This difference... I think is largely differences in BMR estimates used. Spark uses Harris Benedict which always credits me with hundreds of calories more than other formulas. My fitbit bmr is similar to the Mifflin equation. At one point I noticed if I added my exercise minutes to the difference between mifflin and harris benedict bmr estimates that seemed to account for the higher Spark calorie burn.
The final thing that occurred to me, it could be a differential between my fitbit calorie burn and what Spark expects me to burn. Other sites that link to fitbit do this exact thing. Well, in the calorie burn differential report, spark assumes I burn 1685 calories. If I subtract that from the 2058 calories fitbit estimates I burned that day--the result is only 430 calories. So I guess it isn't that either?
Has anyone figured out what this is based on?
| December Minutes: 511