Use the new math.sumprod() function to compute the weighted average.

As compared to fsum(map(mul, data, weights)) the sumprod(data, weights) code is simpler, faster, and more accurate. It is faster because we don't need a succession of calls to mul. It is more accurate because all of the intermediate products are computed losslessly rather than rounded immedately by mul. Also, we no longer need the count() wrapper for this path because sumprod() already incorporates a test to make sure the inputs are the same length.

Baseline timing

% ./python.exe -m timeit -r11 -s 'from random import expovariate as r' -s 'from statistics import fmean' -s 'n=100' -s 'data = [r() for i in range(n)]' -s 'weights = [r() for i in range(n)]' 'fmean(data, weights)'
50000 loops, best of 11: 4.96 usec per loop

Improved timing

% ./python.exe -m timeit -r11 -s 'from random import expovariate as r' -s 'from statistics import fmea
n' -s 'n=100' -s 'data = [r() for i in range(n)]' -s 'weights = [r() for i in range(n)]' 'fmean(data, weights)'
100000 loops, best of 11: 2.06 usec per loop