There are other python packages on github (and elsewhere) for scraping data from fitbit.com, but I had mixed results, especially with "intra-day" data, and especially on heart-rate data. So I wrote this. Data is returned as a DatetimeIndex-ed pandas DataFrame with 15-minute intervals (notably, at least on my Charge HR, heart rate observations are available at 5-minute intervals, but everything else is only at 15-minute intervals, so it returns the median bpm for the three 5-minute interval observations). YMMV.
import pyfitbit
my_data = get_fitbit_data('[email protected]', 'andplzdonotsueme', '2015-08-14', '2015-08-17')
print my_data.loc[my_data.bpm > 0].head()
# output
distance active-minutes bpm floors steps activityLevel calories-burned
dateTime
2015-08-14 11:00:00 0.000000 0 86 0 0 SEDENTARY 21.03480
2015-08-14 11:15:00 0.034443 0 87 0 75 LIGHTLY_ACTIVE 29.09814
2015-08-14 11:30:00 0.122149 0 92 0 266 LIGHTLY_ACTIVE 48.61376
2015-08-14 11:45:00 0.122603 0 92 2 267 LIGHTLY_ACTIVE 53.98932
2015-08-14 12:00:00 0.015149 0 84 0 33 LIGHTLY_ACTIVE 26.99466