Tags give the ability to mark specific points in history as being important
-
1.3
ad1c51f1 · ·Cleaning adapted from Napoly : - M1 : removal of stations with same location (bad set-up from user). - M2 : remove all data using a z-score criteria (at time t, remove all data too far from mean, wich is currently \alpha = 0.05 for the upper tail and \alpha = 0.01 for the lower tail). - M4 : remove stations with corellation too low (currently 0.9) with the median temperature curve. Each step is followed (when necessary, so only M2 here) by an optionnal step descrbed in Napoly : For each station, removes data for any month where the station has too much missing data (if percentage of NaN values > month_missing_threshold = 20%).