Things we have built that other people can use: dataset repositories, interactive tools, and living companions to our surveys. Source code for individual papers is on the code & data page.
fake-news.site
A living companion to our fake news survey (A Survey of Fake News: Fundamental Theories, Detection Methods, and Opportunities, ACM Computing Surveys). It collects what we have learned about the problem in a form you can actually work through.
- Self-guided tutorial — fake news from four angles: the false knowledge it carries, its writing style, how it propagates, and the credibility of its source
- GUIDE research atlas — navigate the literature visually
- Curated papers, datasets, and timely updates
datasets.syr.edu
Our repository of large-scale social network datasets for machine learning, data mining, and computational social science: 19 datasets spanning more than 19 million nodes and 169 million edges across seven kinds of platforms — social networks, blogging, video, photo sharing, music, content, and location-based services. Browsable and sortable by size.
Dataset releases that accompany specific papers — including ReCOVery for COVID-19 news credibility and CHECKED, the first Chinese COVID-19 fake news dataset — are on code & data.
Emotions Dataset
A dataset of user emotions crawled from LiveJournal and hosted here directly, pre-processed so that every result in Emotions in Social Networks: Distributions, Patterns, and Models (Jin and Zafarani, CIKM 2017) can be reproduced. It contains friendships, followers, community memberships, and mood-annotated posts reduced to positive, negative, or neutral polarity, with user and community identifiers anonymised. The dataset page documents each file and gives the citation to use.
WebShapes
Upload a network as an edge list and watch it become a 3D shape. WebShapes samples subgraphs, embeds each one as a point in three dimensions, and fits a shape around the result — a structural fingerprint you can look at and compare. You choose the sampling strategy (six options), the embedding method (four), and the fitting shape (convex hull, cuboid, or sphere), then download the visualization and boundary data.
Built on our ICDM 2018 and KDD 2020 work on interpretable network representations. It makes a good classroom demo — see the tutorial and its recorded talk.
Social Media Mining
The site for our Cambridge University Press textbook: the full text free to read online or download, lecture slides for all ten chapters in PowerPoint and PDF, a table of contents, errata, and citation guidance. Used in 100+ courses across 30+ countries and translated into Chinese and Farsi.
Full details on our textbook & slides page.