NewsTrust Dataset Release: 1.0 (29/07/2015) Data collected from www.newstrust.net and other news sources on July, 2014 _______________________________________________________________________________ Data : http://resources.mpi-inf.mpg.de/impact/credibilityanalysis/data.tar.gz Paper: http://people.mpi-inf.mpg.de/~smukherjee/credibilityanalysis-cikm2015.pdf The data contains 82K NewsTrust stories on 47.6K news articles crawled from 5.7K sources. The data contains NewsTrust member reviews on the articles, and ratings on various qualitative aspects like objectivity, correctness of information, bias and credibility, as well as their interactions with each other. The data has been used in the following paper: Leveraging Joint Interactions For Credibility Analysis In News Communities Subhabrata Mukherjee and Gerhard Weikum In CIKM 2015 If you use this data please cite the paper as follows: BibTex Entry: @InProceedings{mukherjee2014credibilityanalysis, author = {Subhabrata Mukherjee and Gerhard Weikum}, title = {Leveraging Joint Interactions For Credibility Analysis In News Communities}, booktitle = {Proceedings of the 24th ACM International Conference on Information and Knowledge Management (CIKM 2015)}, year = {2015}, } For further questions please contact: smukherjee@mpi-inf.mpg.de _______________________________________________________________________________ ---- FILES ---- - README.txt: This README file. - Each data file is tab-delimited. For each file, the following details are shown below : -- source of the data -- schema -- example record from each file ---- DATA DESCRIPTION AND FORMAT ---- 1. NewsArticles.tsv -- source : 47.6K unique news articles crawled from 5.7K sources like BBC, CNN, National Geographic etc. spanning from 1939 - 2014 -- schema : web-url (tab) article-title (tab) article-body -- example record : http://www.newshounds.us/2006/06/22/sean_hannityand_rick_santorum_cling_to_discredited_finding_of_wmd_in_iraq.php (tab) News Hounds: Rove Tapdances On Subject Of Involvement In McCain Campaign Shake-Up (tab) Reported by Ellen - July 4, 2008 - On last night?s (7/3/08) Hannity & Colmes, Karl Rove was once again brought on to serve as the sole analyst of the Barack Obama campaign ... --------------------- 2. TopicModel.txt (300 topics, 100 words per topic) -- source : JGibbsLDA model ran over the NewsArticle corpus -- schema : Topic Number (newline) < word (space) probability > -- example record : Topic 1: energy 0.012708860562910707 climate 0.012013990189585128 oil 0.00927034728287267 power 0.006507789566571121 water 0.0064570182355688215 change 0.006456022719274659 --------------------- 3. NewsTrustSources.tsv -- source : 668 unique news sources reviewed in NewsTrust -- schema : Attribute (tab) Description - Each new record starts with "web-link" of the source - "overall-ratingLabels" contains the rating attributes, with corresponding values in "overall-ratings" in order - "links" contain hyperlinks inside the given webpage -- example record : web-link http://www.cnn.com/ Expertise Media: Online, Radio, TV Audience: Very large audience (1M or more) (English) Scope: International Political Viewpoint: Neutral viewpoint Owners: Turner Broadcasting System, owned by Time Warner | Managers: Jonathan Klein Related Publications: Anderson Cooper 360, Larry King Live Contact Info: Contact Web Form, Location: Atlanta, GA, US Edited By: Ezra Fox Last Edit: Jul 23, 2006 - 7:29 PM PDT Top Topics: U.S. (477), Politics (391), World (308), Business (192), Sci/Tech (133), Presidential Election 2008 (84) ... overall-ratingLabels [overall, Quality, Facts, Fairness, Information, Insight, Sourcing, Style, Accuracy, Balance, Context, Depth, Enterprise, Expertise, Originality, Relevance, Transparency, Responsibility, Popularity, Recommendation, Credibility, # Reviews, # Views, # Likes, # Emails] overall-ratings [3.4, 3.5, 3.6, 3.5, 3.6, 3.6, 3.3, 3.4, 3.6, 3.1, 3.3, 3.1, 3.1, 3.3, 3.4, 4.0, 3.1, 3.8, 3.3, 3.5, 3.3, 2.4, 4.5, 1.0, 1.0] links [http://www.newstrust.net/feeds/205, http://www.newstrust.net/stories/9755680/edit, http://facebook.com/NewsTrust, http://twitter.com/NewsTrust, http://www.cbsnews.com ...] --------------------- 4. NewsTrustMembers.tsv -- source : Data about 6K NewsTrust members -- schema : Attribute (tab) Description - "overall-ratingLabels" contains the rating attributes, with corresponding values in "overall-ratings" in order - "ratings-given" and "ratings-received" contain the user-user interaction data with the following schema: - ratings(given/received) (tab) rating (tab) (to/from)user (tab) date (tab) on-story-id -- example record : START a-f-james-macarthur http://www.newstrust.net/members/a-f-james-macarthur overall-ratingLabels [Member Level, Activity, Experience, Ratings, Transparency, Validation, Stats, Reviews, Answers, Comments, Ratings Received, Number of Raters, Ratings Given] overall-ratings [3.4, 3.5, 4.6, 3.8, 3.8, 3.0, 26, 224, 8, 4, 3, 18] ratings-received 4 gin-ferrara 04/25/2011 6007608 ratings-received 3 mary-hartney 03/31/2011 5705616 ratings-given 5 kelly-gray 04/30/2011 5695516 ratings-given 4 sara-lindemann 04/28/2011 6007608 ... END --------------------- 5. NewsTrustStories.tsv -- source : 82K NewsTrust stories - each NewsTrust story consists of a news article (in "newsarticle-link") being reviewed by members of the community on different qualitative aspects -- schema : Attribute (tab) Description - Each story starts with its "story-id" - "overall-ratingLabels" contains the overall story rating attributes, with corresponding values in "overall-ratings" in order - each member review of the story has the following schema: - "member-ratingLabels" contains the story rating attributes, with corresponding member assigned ratings in "member-ratings" in order - "member-review" contains the corresponding review with date and user name - "members" contain the user names of the members who reviewed the article in order of their reviews - "links" contain hyperlinks inside the given webpage -- example record : story-id 125621 topic World | U.S. | Politics | Business | Living | Health | United Kingdom | Republicans | Obama Administration | Health Care | Money newsarticle-link http://www.guardian.co.uk/world/2009/aug/11/nhs-united-states-republican-health info The Guardian by Andrew Clark | Aug. 12, 2009 (News Report) source The Guardian written-by Andrew Clark posted-on Aug. 12, 2009 posted-by /members/derek-hawkins overall-ratingLabels [overall, Quality, Facts, Fairness, Sourcing, Style, Context, Depth, Enterprise, Relevance, Popularity, Recommendation, Credibility, # Reviews, # Views, # Likes, # Emails] overall-ratings [3.8, 3.8, 3.8, 4.0, 3.3, 4.0, 4.0, 4.0, 4.0, 4.5, 3.8, 4.1, 3.8, 3.5, 5.0, 1.0, 1.0] member-ratingLabels [overall, Quality, Facts, Fairness, Sourcing, Style, Context, Depth, Enterprise, Relevance, Popularity, Recommendation, Credibility] member-ratings [4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 4.0] member-review Review 4.0 by Glenn LaBauve - Aug. 12, 2009 defends a system rated much better than the US system but vilified by opponents... member-ratingLabels [overall, Quality, Facts, Fairness, Sourcing, Style, Context, Depth, Enterprise, Relevance, Popularity, Recommendation, Credibility] member-ratings [4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 4.0, 3.0, 4.0, 5.0, 4.0, 5.0, 3.0] member-review Review 4.0 by Randy Morrow - Aug. 16, 2009 "The British embassy in Washington is quietly trying to counter inaccuracies... links [http://facebook.com/NewsTrust, http://twitter.com/NewsTrust, http://www.cbsnews.com/...] members [/members/glenn-labauve, /members/randy-morrow, ...] ---------------------