sigmafelix
commented
1 year ago Thank you for making the table in detail. I have a few suggestions:
- Relevant rows need to be classified following "font notes" in README sheet (e.g., derived quantities from NLCD data).
- How about symbolizing instead of colorizing for notes?
- Most spatial resolution units are in areal units (i.e., sq km or sq m), which may confuse users--including me--as spatial resolution is usually expressed as a side length of a pixel in raster datasets. For example, for NLCD derived quantities, some entries include "100 sq m," which is smaller than the actual size of NLCD pixels (30 * 30 = 900 sq m). I understand your choice as areal units would be appropriate for datasets with non-square pixels (e.g., MERRA2). Alternatively, we can just use horizontal and vertical resolutions in that field to cover rectangular pixel cases.
- How about drafting abbreviations for covariate entries (e.g., [category in three letters][measured quantity in up to eight letters][unit of aggregation in zero-padded five digits])? This will help a lot to standardize coding practices when we calculate the covariates.
mitchellmanware
Excel file storing covariates, data sources and relevant metadata is posted on Microsoft Teams.
Teams > NIEHS SET > Group NRT-AP-Model > Files > NRT-AP-Model_Covariates_Metadata.xlsx