Add potential caveat re skewness of distributions

[Nate-Wessel]

One more consideration might be skewness. Maybe the mean is inflated or deflated because of an odd observations. Two ways to look for this might be 1) check if IQR/mean > some cutoff (maybe 1/2) and 2) a more simple check for outliers.

Originally posted by @AdK0101 in #210 (review)

[Nate-Wessel]

I think we could do this simply by checking for major divergence between the median and the mean where the sample is sufficiently high...

In practice, the mean is almost always inflated, not deflated, by a few long travel times / very slow speeds.

[Nate-Wessel]

Some approximate indication of this is given by #268