Modular web toolkit for visualizing news and media sites along different features commonly used for news reliability and objectivity as well as other meta data. Toolkit was demoed at WWW 2018 Conference.
Following are some recent and notable releases of code and data from my group. All remaining code is available at our GitHub account.
713,534 articles from194 news outlets collected between 01/02/2018-30/11/2018, published in ICWSM 2019 Conference containing mainstream sources, satire sources, political commentary sources and sources that have been known to publish unreliable or fake news. In addition, ground truth labels from eight different assessment sites covering multiple dimensions of veracity, including reliability, bias, transparency, and consumer trust.
121K news articles from 92 sources collected in 2017, published in ICWSM 2018 Conference containing mainstream sources, satire sources, political commentary sources and sources that have been known to publish unreliable or fake news.
Code for computing features used for news content analysis, used in NELA Toolkit and many other research publications, especially for studying misinformation.
Given a weighted signed directed network of trust/distrust relations, find a converged balanced network state with the minimal change to the weights in the network. Work published in ACM TWEB.
Refined centrality measures for betweenness and centrality, taking into account multiple paths, measures for local centrality and centraity of a community for networks with a community structure.