Wikipedia Disputed Discussion Corpus (first released on January 2016)
URL: http://www.ccs.neu.edu/home/luwang/

This corpus is distributed together with:

A Piece of My Mind: A Sentiment Analysis Approach for Online Dispute Detection
Lu Wang and Claire Cardie.
Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (ACL), short paper, 2014.


==== Content ====

I. Description of the datasets
II. Contact


==== I. Description of the datasets ====

There are two json files under directory /PATH/TO/dispute_wikipedia:

1) dispute_discussions.json
It contains 3609 disputed discussions collected from Wikipedia talkpages. The description of data collection can be found in our ACL 2014 paper.

2) non_dispute_discussions.json
It contains 3609 discussions sampled from Wikipedia talkpages that are never tagged with disputes by the time this dataset was collected.


>> Data structure

For each file, it contains a list of discussions. Each discussion has fields of 

_page_name: the name of Wikipedia talkpage, e,g, Discussion:1066 Granada massacre
_page_ID: an integer number (unique)
_discussion_name: the discussion title on Wikipedia Talkpage, e.g. Islam and antisemitism
_discussion_ID: an integer number (unique)
_lines: content of the discussion. Each line has fields of "_line_number", "_text", "_user", and "_timestamp". User and timestamp are parsed according to Wikipedia templates (https://en.wikipedia.org/wiki/Wikipedia:Signatures).



==== II. Contact ====

Should you have any questions, please contact luwang@ccs.neu.edu (Lu Wang).