Closed trevorcampbell closed 3 years ago
It should be either in the repository with the experiments' code or in Dropbox or both. I will start tracking it down. Alternatively I think there is some version of it in the invisible institute repository.
I think this issue relates to #7 . In general, I think it would be a good thing to add some supporting datasets, but we need to make sure we apply the same level of rigor as for our main datasets and we need to think carefully about scope.
Another dataset which could fall under this category of "supporting" datasets is the one about crime information per district, as well as the related "historical" work I had done to track down the history of districts numbers over time, to know which districts were merged/split etc.
As an update, I created a new folder for P0-46987
, relating to the original release of unit descriptions by the CPD. I made the "active status" date explicit and did a very minor amount of hand cleaning (the OCR from the PDF failed in a few cases)
Where is the original data listing out all of the names for units (the one I originally built off of when trying to piece together the names of other units)?
Related: