Gap Fill is inaccurate when lots of data is missing.

[ziebelje]

Situation: Have a long period of time where beestat has no data. Say...half a day. With Gap-fill, on a day scale, the other half of the day is going to get the runtime average from the rest of the day. This isn't very accurate.

A quick solution is to disable gap fill for really long durations. A better solution may be to use a metric like runtime/degree day to fill in the gaps. Would have to estimate the missing degree-days is all.