Hi @vagudets,
This development for including the statistical significance of metrics in the scorecards is now ready to be merged. Here is a list of the changes applied:
- New skill metrics are calculated:
rps_syear, rps_clim_syear, crps_syear, crps_clim_syear
. These metrics have an extra dimension:syear
, being each year of the requested data, so no temporal aggregation is applied. - The s2dv functions
RPS.R
andCRPS.R
are temporally included since they have been altered to allow the dimensionsyear
to be returned, by setting the parameterreturn_mean = FALSE
. - Skill metrics are now saved in individual output files rather than all metrics in one file, using the new function:
modules/Saving/R/save_metrics_scorecards.R
. This function uses theCST_SaveExp
function from CSTools. - A new module:
Statistics
has been created, the statistics are also saved in separate output files, in the output folder:outputs/Statistics/
, and with the same filename structure as the skill metrics. - The statistics are also included in the Visualisation module so that when
statistics
is requested inrecipe$Analysis$Workflow$Visualization$plots
, then the statistics are plotted in the same way as the skill metrics. - The calculation of the statistical significance of the metrics is carried out in Scorecards.R, using both skill metrics and statistics.
- The scorecards code has been updated to include the significance of the metrics in the plots, significant values are underlined.
- The SCPlotScorecard function has been replaced by
VizScorecard
, the function from the new ESViz package. - The scorecard functions
.SCTableColors
and.SCLegend
inmodules/Scorecards/R/tmp/Utils.R
have been removed since they are now included in VizScorecard. - The Scorecard.R function has been updated to include all available parameters of the function VizScorecard, so that they can be adjusted by the user. These parameters are also included in the recipe.
Let me know if anything is unclear.
Thanks,
Nadia