Further cleaning and refining

[whanley]

I did a lot of work with OpenRefine to standardize entries in some string fields and find other problems in csv conversion. This is a lot cleaner than previous versions.

I think that the merge conflict is due to some extra error columns being removed?

[shelbygreen]

could you make a new pull request, but uploading the cleaned leon-events file with a different name? we should keep the raw file, but have a "clean" one to work off of

[whanley]

I could upload it under a different name. However the existing leon-events file is already a derivative of the original files (which we received on CDROM). I found problems in that file--extra columns and the like, caused by csv conversion errors--and I think this is a better baseline file.

Other than error correction, I

• standardized spellings of street names
• removed street numbers supplied with street names as well as in the street number column
• standardized names of investigating agencies

I did this in a very conservative way, i.e. I made no judgment other than obvious clustering, using the OpenRefine clustering function and reviewing each decision individually.

[shelbygreen]

Okay, fair! We can use this as the baseline file. I'll fix the merge conflicts and then squash and merge this PR