An interesting tidbit, I decided to try to see what type of cadence signature I could find for Mimi even with her garbage cadence data. I've just analyzed her Sept 10, 2017 Morning data for a hill grade from 0-1% by plotting the speed on the Y-axis and cadence in spm on the X-axis. Thus, this graph gives her average step length as the slope of the straight regression line. I'm using a 10 second rolling average for the GPS data that calculates the speed because of the inherent noise in GPS data. The cadence data I leave untouched.
So far, this first file says Mimi has an average step length of 80.5 cm or about 31.7 inches with a R^2 of 0.6929. With this graph, you can tell there is a slow run speed that centers around 150 spm and a fast run speed centred around 180 spm. I would prefer a higher R^2 value but there is inherent noise with GPS and cadence measurements.
There is nothing conclusive about this analysis yet, it's more for curiosity at this point until I go through more of them.