There are number of hours in RTS_Data/timeseries_data_files/WIND/DAY_AHEAD_WIND.csv
that have unreasonably high forecast errors. The worst two hours are
Date, Time, Forecast, Observed total (from REAL_TIME_wind.csv)
2020-02-27, 19:00:00, 0.1, 1918.03333333333
2020-03-05, 19:00:00, 0.1, 1918.03333333333
The fact that it is for the same hour of the day with the same unrealistic values but on different days makes me suspect a bug in the software that created the files. Note that these forecasts
differ from neighboring hours by more than two orders of magnitude (and the neighboring
hours are also quite bad).
There are number of hours in RTS_Data/timeseries_data_files/WIND/DAY_AHEAD_WIND.csv that have unreasonably high forecast errors. The worst two hours are
Date, Time, Forecast, Observed total (from REAL_TIME_wind.csv) 2020-02-27, 19:00:00, 0.1, 1918.03333333333
2020-03-05, 19:00:00, 0.1, 1918.03333333333
The fact that it is for the same hour of the day with the same unrealistic values but on different days makes me suspect a bug in the software that created the files. Note that these forecasts differ from neighboring hours by more than two orders of magnitude (and the neighboring hours are also quite bad).