Grouping based on PageURL sometimes groups multiple datasets together e.g. http://www.moray.gov.uk/moray_standard/page_110140.html actually has 8 datasets but they all collapse down to 1 with my code (and just get “Car Parks” metadata. Maybe use (hash of) more fields as key to the dict?
Have now added title to the key and that looks to have improved things.
So many NaNs in the data…
Should swap them out for something else either in the numpy array or when building Dataset objects from it.
Ends up sorted to top of the list
I think created and updated dates are the wrong way round for entries from CKAN - regularly have updated before created, or an updated date but no created.
I’m rendering all the descriptions as if they are markdown but that’s not the case for all of them. Might be mangling some. Or there might be a more appropriate rendering.