farewe / invacost Goto Github PK

I have added an error check for missing values when expandYearlyCosts is used. In doing this I found that only "Probable_starting_year_low_margin" and "Probable_ending_year_low_margin" have complete data. "Probable_...year" and "Probable..._year_high_margin" do not. Not sure if this should be part of the response to the user (i.e. explain why the data are missing).

Suggestion from reviewer in MEE: include accuracy assesments in model plots.
What should we include? RMSE? RMSE on calibration points only?

So far it is the only metric we have included as a goodness-of-fit metric is the RMSE, should we implement new ones? e.g. the median absolute-deviation?

With the way you have implemented 'sandwich' on the least squares regression would it be more appropriate to call this something different on the plots and tables? "Weighted least squares regression"

I cleaned up some errors here, specifically with regards to the returned lists. I also found the Note regarding "rawAvgCosts()" to be confusing, but wasn't sure how to adjust appropriately.
I had one additional comment, and that is that NAs are returned from "calculateAvgRawCosts()" when there is not data in a year range. This seems slightly at odds with the documentation which says: "Missing data for specific years will be considered as zero." (in Details).
No error reporting appeared necessary.