B) The same algorithm was used to fill in the data from 06/25/18 through 06/30/18.
C) Data for 2017 was recorded by a LabJack USB data logger. It's numbers seem to be dramatically higher than the subsequent Vantage Vue weather station, and nearby stations don't show a similar change between those two years, thus it seems appropriate to discount that years data.
D) Data for 2019 was very incomplete, no attempt has been made to interpolate.
E) Some data for 2020 was obviously bad, no attempt has been made to
interpolate. As an example, there were summer days where the
diurnal difference (nighttime low vs. daytime high) indicated less
than a degree, even though in reality the temps were in the 50's
vs. 80's range. Another example is that both the high and low for
26-Mar through 15-Apr was 52.3 - every 15 mins for 3 weeks, current,
high, and low temps were unchanged.