Open imarkonis opened 3 months ago
Added two trend functions that use TheilSen.
The first trends_lon_lat_boot parallizes across lon, lat and calculates a trend for each lon, lat grid. The user has to define a "pollutant" (wording from openair).
The second trends_dataset_boot also parallizes across lon, lat BUT this calculates a trend for each lon, lat AND dataset. The user has to define a "pollutant" (wording from openair).
The openair package requires the time to be in a specific wording. Should I adapt the functions to test the time has the correct name, wording?
See example
evap_datasets[, year := as.numeric(as.character(year))]
evap_datasets[, date := paste0(year, "-01-01 00:00:00")]
evap_datasets[, date := as.POSIXct(date)]
Let's keep it as it is for now, but with a comment on openair date format requirement
Copy trend estimation functions from evap_trend to geo_functions.