Outlier detection on graphs? #13
Description
If we get a really wacky sensor reading (right now @ CIEE, one of the IMT550C's has reported -2e8 and 2e8 as temperatures), we shouldn't plot it because it ruins the y-scale of the sparklines. Can we have a bound on the temperature plots? This is linked to the issue of "sensors shouldn't give bad readings in the first place"