249 datasets tagged as the platform “Unaffiliated”.
249 datasets — page 2 of 5.
A Catster dataset of 149,700 users, 5,449,275 friendships, collected by Catster website (catster.com). Each record contains undirected edges between Catster users representing friendships among cat owners and cat lovers. Catster social network capturing friendships in a niche social networking site for cat owners, cat lovers, and pet enthusiasts. Suitable for social network analysis, community detection, niche social media studies, pet owner behavior analysis.
A Petster.com (pet social network — cats) dataset of 149,700 nodes and 5,449,275 edges, collected by KONECT via Petster crawl. Each record contains directed: False, weighted: False, signed: False. Friendship network of cat profiles on a pet social networking site.. Suitable for social network analysis, community detection.
A Princeton WordNet lexical database dataset of 146,005 nodes and 656,999 edges, collected by Princeton University. Each record contains directed: False, weighted: False. Semantic network of English words linked by synonymy, hypernymy, and other lexical relations from WordNet.. Suitable for natural language processing, semantic network analysis, word embedding evaluation.
A ReCOVery (news credibility repository) dataset of 2,029 news articles with reliability labels and social media engagement, collected by Zhou, Mulay, Ferrara, and Zafarani via news websites and Twitter (2020). Each record contains news article title and body, reliability label (1=reliable, 0=unreliable), publisher, political bias, country, author, and social media engagement data. The dataset is notable for its multimodal repository combining news content with social media engagement, designed for COVID-19 news credibility research. Intended for research into fake news detection, credibility assessment, political bias analysis, and multimodal misinformation research.
A Scored.co (c/Christianity) dataset of 10,065 submissions and 114,801 comments from 699 unique authors in c/Christianity (2020-2023), collected by iDRAMA Lab via Scored.co API (2024). Each record contains post/comment text, author, score up/down, NSFW/deleted/twitter-crosspost flags, date, UUID. c/Christianity on Scored.co — Christian religious discussion community. 124,866 total posts from 699 authors over 4 years. Suitable for deplatforming research, community behavior analysis, alternative platform discourse, content moderation effects, longitudinal community evolution.
A Scored.co (c/Shithole) dataset of 25,200 submissions and 73,520 comments from 804 unique authors in c/Shithole (2020-2023), collected by iDRAMA Lab via Scored.co API (2024). Each record contains post/comment text, author, score up/down, NSFW/deleted/twitter-crosspost flags, date, UUID. c/Shithole on Scored.co — General unmoderated discussion. 98,720 total posts from 804 authors over 4 years. Suitable for deplatforming research, community behavior analysis, alternative platform discourse, content moderation effects, longitudinal community evolution.
A BlogCatalog (v2) dataset of 97,884 bloggers and 1,668,698 friendship edges, collected by Agarwal et al. / Tang & Liu via BlogCatalog crawl. Each record contains blogger-blogger friendship edges, user IDs. The dataset is notable for its blogosphere friendship network from BlogCatalog, used for familiar stranger detection and latent social dimensions research. Intended for research into social network analysis, community detection, group membership prediction, and blogosphere structure studies.
A Prosper.com (peer-to-peer lending platform) dataset of 89,269 nodes and 3,330,225 edges, collected by KONECT via Prosper API. Each record contains directed: True, weighted: True, signed: False, temporal: True. Directed trust network from a peer-to-peer lending platform where users vouch for borrowers. Temporal edges with Unix timestamps enable longitudinal trust propagation analysis.. Suitable for trust network analysis, temporal network analysis, peer-to-peer lending.
A BlogCatalog (blogging social network) dataset of 88,784 bloggers and 2,093,195 friendship edges, collected by Agarwal et al. via BlogCatalog crawl (July 2009). Each record contains blogger-blogger friendship edges and user IDs. The dataset is notable for its classic blogosphere friendship network used for familiar stranger detection and social identity research. Intended for research into social network analysis, community detection, familiar stranger identification, and blogosphere structure studies.
A News Articles (multiple sources) dataset of 85,917 articles (43,610 fake, 34,941 true, 7,366 russian propaganda), collected by Steven Peutz via Kaggle (2022); Russian propaganda subset sourced from the EUvsDisinfo Project (2015-present). Each record contains full article text and veracity label (fake, true, or russian_propaganda). combination of general fake/true news articles with a dedicated Russian propaganda subset from the EUvsDisinfo database, enabling comparative analysis across disinformation types. Suitable for fake news detection, misinformation topic modeling, propaganda characterization, and classification benchmarking.
A Flickr (photo-sharing) dataset of 80,513 users and 5,899,882 friendship edges, collected by Tang & Liu via Flickr crawl (2009). Each record contains user-user friendship edges, group memberships (195 groups), and user IDs. The dataset is notable for its classic photo-sharing social network with explicit group memberships, used as a benchmark for latent social dimensions. Intended for research into social network analysis, community detection, group membership prediction, and photo-sharing platform structure studies.
A Jester Online Joke Recommender System (UC Berkeley) dataset of 73,521 nodes and 4,136,360 edges, collected by K. Goldberg et al. via Jester system. Each record contains directed: False, weighted: True, temporal: False. Bipartite user-joke rating network with continuous ratings from -10 to +10 — one of the earliest large-scale collaborative filtering datasets.. Suitable for recommender systems, collaborative filtering, continuous rating prediction.
A TextDetox 2024/2025 shared task (15 languages) dataset of 71,374 posts across 15 languages with binary toxicity labels, collected by Dementieva et al. for PAN/CLEF TextDetox 2024. Each record contains post text, binary toxicity label (0=non-toxic, 1=toxic), language. Combined multilingual toxicity benchmark covering 15 languages including low-resource languages (Amharic, Tatar, Hindglish) — 71,374 balanced samples total. Suitable for multilingual toxicity detection, cross-lingual transfer learning, low-resource NLP, content moderation benchmarking.
A Scored.co (c/WSBets) dataset of 9,370 submissions and 56,988 comments from 2,042 unique authors in c/WSBets (2020-2023), collected by iDRAMA Lab via Scored.co API (2024). Each record contains post/comment text, author, score up/down, NSFW/deleted/twitter-crosspost flags, date, UUID. c/WSBets on Scored.co — WallStreetBets-style financial discussion. 66,358 total posts from 2,042 authors over 4 years. Suitable for deplatforming research, community behavior analysis, alternative platform discourse, content moderation effects, longitudinal community evolution.
A Usenet (sample) dataset of 64,999 posts from 4754 newsgroups, collected by OwnedByDanes via Hugging Face. Each record contains post text, newsgroup, date, author, subject. Sample of Usenet posts from 1980–2013, with 5,000 posts per top‑level hierarchy. Covers comp.*, sci.*, rec.*, alt.*, etc.. Suitable for historical discourse analysis, early internet culture, topic evolution, authorship attribution.
A Gnutella (August 31, 2002) dataset of 62,586 nodes and 147,892 edges, collected by Ripeanu et al. via Gnutella crawl (2002). Each record contains directed peer-to-peer connections in the Gnutella network. The dataset is notable for its snapshot from August 31, 2002, part of a temporal sequence of 9 snapshots. Intended for research into peer-to-peer network analysis, temporal graph evolution, and network resilience studies.
A ISOT (fake news dataset) dataset of 44,898 news articles labeled fake/real, collected by ISOT via Reuters and various sources (2017). Each record contains news article text (title + content) and binary label (is_fake: 1=fake, 0=real). The dataset is notable for its well-balanced dataset of fake and real news articles for misinformation research. Intended for research into fake news detection, text classification, and misinformation analysis.
A Academic co-authorship bipartite network dataset of 38,741 nodes and 58,595 edges, collected by T. Opsahl. Each record contains directed: False, weighted: False, signed: False. Large bipartite collaboration network of authors and papers — useful for studying academic collaboration patterns.. Suitable for bipartite network analysis, academic collaboration, community detection.
A Scored.co (c/News) dataset of 18,801 submissions and 18,518 comments from 647 unique authors in c/News (2020-2023), collected by iDRAMA Lab via Scored.co API (2024). Each record contains post/comment text, author, score up/down, NSFW/deleted/twitter-crosspost flags, date, UUID. c/News on Scored.co — General news discussion. 37,319 total posts from 647 authors over 4 years. Suitable for deplatforming research, community behavior analysis, alternative platform discourse, content moderation effects, longitudinal community evolution.
A Gnutella (August 30, 2002) dataset of 36,682 nodes and 88,328 edges, collected by Ripeanu et al. via Gnutella crawl (2002). Each record contains directed peer-to-peer connections in the Gnutella network. The dataset is notable for its snapshot from August 30, 2002, part of a temporal sequence of 9 snapshots. Intended for research into peer-to-peer network analysis, temporal graph evolution, and network resilience studies.
A Scored.co (c/Memes) dataset of 8,344 submissions and 25,287 comments from 707 unique authors in c/Memes (2020-2023), collected by iDRAMA Lab via Scored.co API (2024). Each record contains post/comment text, author, score up/down, NSFW/deleted/twitter-crosspost flags, date, UUID. c/Memes on Scored.co — General meme sharing community. 33,631 total posts from 707 authors over 4 years. Suitable for deplatforming research, community behavior analysis, alternative platform discourse, content moderation effects, longitudinal community evolution.
A Manosphere forum (Love-Shy) dataset of 29,232 posts from 826 unique authors on Love-Shy, collected by Ribeiro et al. via web crawl (2021). Each record contains post text (text_post), author (author), timestamp, post metadata. 29,232 posts from Love-Shy — part of the Evolution of the Manosphere dataset covering incel, MGTOW, PUA, and related anti-feminist communities across 9 forums. Suitable for manosphere discourse analysis, incel radicalization research, hate speech detection, community evolution, misogynistic content analysis, toxicity measurement.
Contacts between people measured by small wireless devices they carried, from the Haggle project (Chaintreau et al.), as distributed by KONECT (contact): 28,244 timestamped contacts among 274 people, forming 2,124 distinct pairs. An edge means two people's devices were within range of each other; timestamps are relative and the records span about four days. Useful for opportunistic networking, temporal network analysis and epidemic modelling on human contact data.
The co-authorship network of the arXiv High Energy Physics – Phenomenology (hep-ph) section, as distributed by KONECT (ca-cit-HepPh), from SNAP's arXiv data: 28,093 authors linked by 4,596,803 collaboration records, one for each pair of authors on a paper, published between March 1992 and March 2002. The records form 3,148,447 distinct co-author pairs, each timestamped with its paper's date. Useful for studying scientific collaboration, the growth and densification of networks over time, and community structure.
A arXiv High Energy Physics — Theory (1991–2003) dataset of 27,770 nodes and 352,807 edges, collected by J. Leskovec et al. via arXiv. Each record contains directed: True, weighted: False. Citation network of HEP-TH papers on arXiv — directed, covers theoretical physics literature.. Suitable for citation network analysis, academic influence, graph mining.
A Gnutella (August 24, 2002) dataset of 26,518 nodes and 65,369 edges, collected by Ripeanu et al. via Gnutella crawl (2002). Each record contains directed peer-to-peer connections in the Gnutella network. The dataset is notable for its snapshot from August 24, 2002, part of a temporal sequence of 9 snapshots. Intended for research into peer-to-peer network analysis, temporal graph evolution, and network resilience studies.
A Internet autonomous system topology (CAIDA, 2007) dataset of 26,475 nodes and 53,381 edges, collected by CAIDA via BGP routing data. Each record contains directed: False, weighted: False. Internet autonomous system (AS) peering network — nodes are AS networks, edges are peering relationships.. Suitable for internet topology analysis, network robustness, community detection.
A Math Overflow (ALL) dataset of 24,818 users and 239,978 temporal edges, collected by Paranjape, Benson, Leskovec via Stack Exchange Data Dump (2017). Each record contains directed temporal edges (all) between users with Unix timestamps. The dataset is notable for its ALL interaction type from a 6.4-year temporal network of academic Q&A. Intended for research into temporal network analysis, motif discovery, social interaction dynamics, and academic community analysis.
A FakeNewsNet (fake news with social context) dataset of 23,196 news articles labeled fake/real with social media context, collected by Shu et al. via Politifact and GossipCop (2019). Each record contains news title, article text, binary label (0=real, 1=fake), source (politifact/gossipcop), publish date, and URL. The dataset is notable for its dataset with social context and spatiotemporal information for studying fake news propagation. Intended for research into fake news detection, social media analysis, misinformation propagation, and multimodal research.
A Cora machine learning paper database dataset of 23,166 nodes and 91,500 edges, collected by McCallum et al.. Each record contains directed: True, weighted: False. Classic citation network of machine learning papers — one of the most used benchmarks for graph neural network node classification.. Suitable for graph neural networks, node classification, citation network analysis.
A Edinburgh Associative Thesaurus (word association study) dataset of 23,132 nodes and 312,310 edges, collected by G.R. Kiss et al.. Each record contains directed: True, weighted: True. Directed word association network from the Edinburgh Associative Thesaurus — edges connect stimulus words to response words.. Suitable for semantic network analysis, word association modeling, NLP.
A Gnutella (August 25, 2002) dataset of 22,687 nodes and 54,705 edges, collected by Ripeanu et al. via Gnutella crawl (2002). Each record contains directed peer-to-peer connections in the Gnutella network. The dataset is notable for its snapshot from August 25, 2002, part of a temporal sequence of 9 snapshots. Intended for research into peer-to-peer network analysis, temporal graph evolution, and network resilience studies.
A CLEF CheckThat! 2024 shared task (english) dataset of 22,501 sentences with binary check-worthiness labels (24.1% check-worthy), collected by CLEF CheckThat! Lab organizers (2024). Each record contains sentence text, binary check-worthiness label (Yes/No), sentence ID, source document. CLEF 2024 CheckThat! Task 1 (english) provides 22,501 sentences from political debates and social media annotated for check-worthiness — whether the claim merits fact-checking. Suitable for check-worthiness detection, fact-checking automation, misinformation research, NLP classification benchmarks, political discourse analysis.
A Math Overflow (A2Q) dataset of 21,688 users and 90,489 temporal edges, collected by Paranjape, Benson, Leskovec via Stack Exchange Data Dump (2017). Each record contains directed temporal edges (a2q) between users with Unix timestamps. The dataset is notable for its A2Q interaction type from a 6.4-year temporal network of academic Q&A. Intended for research into temporal network analysis, motif discovery, social interaction dynamics, and academic community analysis.
A CLEF CheckThat! 2024 shared task (spanish) dataset of 19,948 sentences with binary check-worthiness labels (15.7% check-worthy), collected by CLEF CheckThat! Lab organizers (2024). Each record contains sentence text, binary check-worthiness label (Yes/No), sentence ID, source document. CLEF 2024 CheckThat! Task 1 (spanish) provides 19,948 sentences from political debates and social media annotated for check-worthiness — whether the claim merits fact-checking. Suitable for check-worthiness detection, fact-checking automation, misinformation research, NLP classification benchmarks, political discourse analysis.
A PubMed dataset of 19,717 nodes, 44,327 edges, 3 classes, collected by Yang et al. (ICML 2016). Each record contains node features, citation edges, class labels. PubMed network with bag-of-words features. Suitable for node classification, semi-supervised learning, graph neural networks.
A arXiv Astrophysics (collaboration network 1992–2003) dataset of 18,771 nodes and 198,050 edges, collected by J. Leskovec et al. via arXiv. Each record contains directed: False, weighted: False. Co-authorship network of astrophysics papers on arXiv — nodes are authors, edges connect co-authors.. Suitable for collaboration network analysis, community detection, academic social networks.
Yelp coffee-shop reviews for computational branding analytics (Wang, Lin & Kominek, EMNLP 2013): 17,894 reviews in train (10,495), dev (3,617) and test (3,782) splits. Each row holds the brand (Starbucks, Dunkin', other), satisfaction (negative, positive, neutral), the reviewer's gender (missing as '?' in some rows), US region, and the review text. There are no review, business or date fields. Research use only, per the README.
A Math Overflow (C2Q) dataset of 16,836 users and 101,329 temporal edges, collected by Paranjape, Benson, Leskovec via Stack Exchange Data Dump (2017). Each record contains directed temporal edges (c2q) between users with Unix timestamps. The dataset is notable for its C2Q interaction type from a 6.4-year temporal network of academic Q&A. Intended for research into temporal network analysis, motif discovery, social interaction dynamics, and academic community analysis.
A Escort review website (sex worker review community) dataset of 16,730 nodes and 39,044 edges, collected by Rocha et al. via web crawl. Each record contains directed: False, weighted: True, signed: True, temporal: True. Bipartite signed network of escort reviews: clients rate escorts +1 (positive), 0 (neutral), or -1 (negative), with Unix timestamps. One of the few large-scale signed bipartite networks available.. Suitable for signed network analysis, bipartite network analysis, trust propagation.
A Math Overflow (C2A) dataset of 13,840 users and 81,121 temporal edges, collected by Paranjape, Benson, Leskovec via Stack Exchange Data Dump (2017). Each record contains directed temporal edges (c2a) between users with Unix timestamps. The dataset is notable for its C2A interaction type from a 6.4-year temporal network of academic Q&A. Intended for research into temporal network analysis, motif discovery, social interaction dynamics, and academic community analysis.
A Free On-line Dictionary of Computing (FOLDOC) dataset of 13,356 nodes and 120,238 edges, collected by Denis Howe. Each record contains directed: True, weighted: True. Directed network of computing terms linked by cross-references in the FOLDOC dictionary.. Suitable for knowledge network analysis, dictionary network structure, NLP.
A LIAR (fact-checking benchmark) dataset of 12,789 labeled claim statements from Politifact, collected by Wang via Politifact fact-checking (2017). Each record contains claim text, 6-class veracity label, speaker metadata (party, state, job title), credit history counts (barely_true/false/half_true/mostly_true/pants_fire), and context. The dataset is notable for its benchmark dataset for fake news detection with fine-grained 6-class labels, rich speaker metadata, and credit history. Intended for research into fake news detection, fact-checking, political discourse analysis, and misinformation research.
A Anybeat dataset of 12,645 users, 49,132 connections, collected by Fire et al. (2012) via Anybeat platform. Each record contains undirected edges between Anybeat users representing social connections in the community platform. Anybeat social network capturing user connections in a public gathering place where people interact with others from around their neighborhood or across the world. Suitable for link prediction, social network analysis, community detection, online community behavior studies.
A LASAGNE online social network dataset dataset of 12,643 nodes and 57,453 edges, collected by KONECT. Each record contains directed: True, weighted: True. Directed network from the LASAGNE project — a small online social network dataset.. Suitable for social network analysis, community detection.
A DBLP Computer Science Bibliography dataset of 12,591 nodes and 49,743 edges, collected by KONECT via DBLP database. Each record contains directed: True, weighted: False. Directed citation network of computer science papers in the DBLP bibliography.. Suitable for citation network analysis, academic influence, community detection.
A CLEF CheckThat! 2025 Task 2 (english) dataset of 11,374 social media post and normalized claim pairs, collected by CLEF CheckThat! Lab organizers (2025). Each record contains social media post text containing a factual claim, normalized canonical claim statement. CLEF 2025 CheckThat! Task 2 (english) provides 11,374 pairs of social media posts and their normalized factual claims, enabling claim normalization and canonicalization research. Suitable for claim normalization, fact-checking automation, NLP text generation, multilingual claim extraction, social media fact verification.
A Gnutella (August 04, 2002) dataset of 10,876 nodes and 39,994 edges, collected by Ripeanu et al. via Gnutella crawl (2002). Each record contains directed peer-to-peer connections in the Gnutella network. The dataset is notable for its snapshot from August 04, 2002, part of a temporal sequence of 9 snapshots. Intended for research into peer-to-peer network analysis, temporal graph evolution, and network resilience studies.
A PGP (Pretty Good Privacy) key signing network dataset of 10,680 nodes and 24,316 edges, collected by M. Boguñá et al.. Each record contains directed: False, weighted: False, signed: False. Web of trust network from PGP key signings — nodes are PGP users, edges are key signatures.. Suitable for trust network analysis, community detection, graph algorithm benchmarking.
A TUAW (The Unofficial Apple Weblog) dataset of 10,485 blog posts, collected by Agarwal, Tang, Liu, and Yu via TUAW crawl (2004-2007). Each record contains blog post title, text, author, categories, publication date, inlinks, outlinks, and comment counts. The dataset is notable for its blog dataset with rich metadata, representing a closed community of Apple-focused bloggers. Intended for research into influential blogger identification, blog post analysis, text classification, and influence modeling.
Petster
Twitter (X)
Scored
BlogCatalog
Flickr
Usenet
Gnutella
The internet
Stack Exchange
Yelp