Human-evaluated data for supervised learning

[austinbrian]

This is a csv of evaluated SERVICENOTES chosen at random. The "orig_index" column is the index number of the original data file. This can be used for evaluating LDA topics and other machine learning methods.

[jasonasher]

Thanks for submitting this. I'm in the process of moving our repository to the Code for DC organization account. Would you mind submitting it there?
https://github.com/codefordc/the-rat-hack/

Also, I wasn't sure what to make of some of the service note entries - see, e.g. index number 2995352 in your submission. Is that an inspector making that statement, or the original complaint text? In the former it's a 'y', and in the latter a 'u', I would say. Right?

[supermdat]

Thanks for this, @austinbrian!

I have a quick question about something I think I misunderstood. What does "original_index" in your dataset map to in the original file? I thought the "index" in the original file was "SERVICEREQUESTID"? But the format of "original_index" looks a bit different than that of "SERVICEREQUESTID".