All of the files and scripts associated with 
re-annotating classic classification data with 
multiple annotations from crowdflower.

input/  # scripts and data that were used to run the user study
  20ng-reannotate.csv contains the actual 1000 files that were re-annotated
  
cml  # the precise gui that annotators used

results/  # raw annotations gotten from crowdflower (in various forms)

dataset/ # scripts to produce a dataset from data in the results folder
  cfgroups1000.json # the main dataset created from this project. If you take only one thing from this repo, it should be this.