Directed network

93 datasets whose data is: directed network. Computed from the tables of each dataset's ECHO Data Package.

93 datasets.

Twitter Follow Graph, August 2009 (MPI-SWS, ICWSM 2010) Analyzed

The follow graph of Twitter shared by the Max Planck Institute for Software Systems (MPI-SWS) for the study "Measuring User Influence in Twitter: The Million Follower Fallacy" (ICWSM 2010): 1,963,263,821 directed follow links among 52,579,682 accounts, with account IDs anonymised, so the file carries structure only, and a table of the number of new adopters per day of seven retweeting conventions (RT, via, Retweeting, Retweet, HT, R/T and the recycling symbol) over 883 days. 43,983,853 accounts follow at least one account and 49,395,940 are followed by at least one. The page that publishes the data says it is based on the Twitter network snapshot of September 2009 and that the tweet information is not shared.

52.6M records52.6M nodes2B edges2010

Twitter 2010 Follower Network Analyzed

The Twitter follower graph crawled in 2010 by Kwak, Lee, Park and Moon, as the Laboratory for Web Algorithms (LAW) of the University of Milan publishes it as twitter-2010: 1,468,365,182 arcs among 41,652,230 accounts, with node numbers renumbered contiguously from 0 and a table that maps each node number back to the numeric Twitter user ID. LAW defines an arc from x to y as 'y is a follower of x'.

41.7M records41.7M nodes1.5B edges2010

Gab Posts Dataset Analyzed

A Gab (alt-right social network) dataset of 22,112,812 Gab posts from January 2018 (13,871,421 original posts, 8,241,391 reposts), collected by Zannettou et al. via Gab API (2018). Each record contains post body text, username, verified flag, published_at, post type (post/repost), topic category, language, likes/reposts/replies counts, sentiment score, isPopular. 22,112,812 Gab posts from January 2018 — one of the earliest large-scale Gab datasets, capturing the platform during its growth phase as an alt-right Twitter alternative. Suitable for alt-right discourse analysis, fringe platform research, meme propagation, hate speech detection, cross-platform information flow, deplatforming research.

22.1M records2019

LiveJournal Emotions Dataset (Jin & Zafarani 2017) Analyzed

12,404,868 LiveJournal posts with the mood their authors selected, labelled negative (5,014,041), neutral (2,894,738) or positive (4,496,089), by 115,444 users, with the time of each post; the friendship links (246,164) and follower links (793,948) among the users; and the 2,473,074 memberships of 84,925 users in 200,208 communities. The data were crawled from LiveJournal within three hops of a starting user and preprocessed by Shengmin Jin and Reza Zafarani for their paper 'Emotions in Social Networks: Distributions, Patterns, and Models' (CIKM 2017); users and communities are identified only by numbers assigned by the authors.

12.4M records106.6K nodes961.4K edges2017

Wikipedia Hyperlinks (English) Analyzed

Wikipedia links, English network, a network of the Koblenz Network Collection (KONECT) (Article–article wikilink). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 378,142,420 links among 12,150,976 nodes.

12.2M records12.2M nodes378.1M edges2013

Twitter Social Network (Syracuse Data Lab) Analyzed

A follower network of Twitter users, as distributed by the Data Lab's Social Computing Data Repository: a directed link from a user to a user they follow (the readme: user 1 following user 2 is written 1,2). The readme describes Twitter as a social news website and names neither who collected the data nor when. This copy holds 85,331,845 links among 11,316,811 users and a list of 11,316,811 node numbers.

11.3M records11.3M nodes85.3M edges2026

YouTube SFU Crawls 2007-2008 (Collection) Limited

Repeated crawls of YouTube made in 2007 and 2008 by Xu Cheng, Cameron Dale and Jiangchuan Liu of Simon Fraser University, as published on the data page of the university's Network Systems Group. There are 94 breadth-first crawls of YouTube's lists of related videos, 35 started between 22 February and 18 May 2007 and 59 between 27 March and 28 July 2008. Each starts from a set of seed videos and follows each video's list of related videos down to depth 3 (to depth 4 in the first crawl, to depth 2 in 2007-0302). The package holds one record for each video a crawl visited (8,490,464 records; uploader, age, category, length, number of views, average rating, number of ratings and number of comments) and one directed link for each entry in a visited video's list of related videos (129,437,347 links, with the position in the list as rank), all marked with the crawl they come from. It also holds 28 update crawls (4,623,189 records with the views, rating, number of ratings and number of comments of videos already seen; 7 in 2007 and 21 in 2008), the video file sizes and bitrates measured in three further crawls (552,647 records), and a table that describes each of the 126 crawls from its log.

8.5M records2008

Twitter Events — All Events (Cumulative) Analyzed

Multiplex networks of Twitter users during 5 events of 2013 and 2014 (Cannes2013, MLKing2013, MoscowAthletics2013, NYClimateMarch2014, ObamaInIsrael2013), as collected by Elisa Omodei, Manlio De Domenico and Alex Arenas and distributed by Netzschleuder. A node is a user, identified only by an anonymous number; a directed link from i to j means that user i retweeted, mentioned or replied to user j during the event, weighted by the number of such interactions. The package holds 3,238,746 nodes and 6,014,230 links (2,921,236 retweets, 2,383,605 mentions, 709,389 replies) of 5 networks; the column layer names the event, so a single event is a filter of each table.

5.2M records3.2M nodes5.2M edges2015

LiveJournal Social Network Analyzed

LiveJournal network, a network of the Koblenz Network Collection (KONECT) (User–user friendship). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 68,475,391 links among 4,846,609 nodes.

4.8M records4.8M nodes69M edges2008

Wikipedia Hyperlinks (German) Analyzed

Wikipedia links, German network, a network of the Koblenz Network Collection (KONECT) (Article–article wikilink). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 81,626,917 links among 3,225,565 nodes.

3.2M records3.2M nodes81.6M edges2013

Wikipedia Hyperlinks (French) Analyzed

Wikipedia links, French network, a network of the Koblenz Network Collection (KONECT) (Article–article wikilink). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 102,382,410 links among 3,023,165 nodes.

3M records3M nodes102.4M edges2013

Twitter 2010: URL Tweets and Follower Graph (Lerman, USC ISI) Analyzed

Tweets that contain a URL, collected through Twitter search in the second half of September 2010 and the first day of October 2010, with the follower graph among the users who posted them, released by Kristina Lerman (USC Information Sciences Institute) for her studies of how information spreads on Twitter. The package holds 2,859,764 rows linking a URL to a tweet (2,789,194 distinct tweets, 66,064 distinct URLs, 736,930 users), the list of those 736,930 users with their follower and followee counts, and 36,743,448 follower links among them. The tweet text is not part of the release: each row has the tweet ID, the posting user, the time, the posting client, the reply target and the URL.

2.9M records700K nodes36.7M edges2012

Wikipedia Hyperlinks (Russian) Analyzed

Wikipedia links, Russian network, a network of the Koblenz Network Collection (KONECT) (Article–article wikilink). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 82,056,101 links among 2,853,118 nodes.

2.9M records2.9M nodes82.1M edges2013

Flixster Friendship Network (December 2010) Analyzed

Flixster is a social movie site that lets users share movie ratings, discover movies and meet others with similar taste. This is the friendship network among its users, crawled in December 2010 by Javier Parra, as distributed by the Data Lab's Social Computing Data Repository. This copy holds 9,197,337 friendship links among 2,523,386 users and a list of 2,523,386 user numbers.

2.5M records2.5M nodes9.2M edges2010

LiveJournal Friendship Network (July 2010, Syracuse Data Lab) Analyzed

The LiveJournal friendship network of the Data Lab Social Computing Data Repository (Syracuse University): LiveJournal, a virtual community where Internet users keep a blog, journal or diary. The network was crawled in July 2010 by Reza Zafarani, William D. Cole, Huan Liu. This copy holds 14,608,137 directed links, each from a user to a friend of that user (the readme: the line 1,2 means user 1 is friend with user 2), among 2,238,731 distinct users, and a list of 2,238,731 user numbers (nodes.csv). Users are numbers; the files hold no names or profile data.

2.2M records2.2M nodes12.8M edges2010

Wikipedia Hyperlink Growth Network Analyzed

Wikipedia, English network, a network of the Koblenz Network Collection (KONECT) (Article–article links). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 39,953,145 links among 1,870,709 nodes, timed from 19 February 2001 to 5 April 2007.

1.9M records1.9M nodes40M edges2009

Wikipedia Hyperlinks (Italian) Analyzed

Wikipedia links, Italian network, a network of the Koblenz Network Collection (KONECT) (Article–article wikilink). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 91,555,008 links among 1,865,965 nodes.

1.9M records1.9M nodes91.6M edges2013

Wikipedia Hyperlinks (Japanese) Analyzed

Wikipedia links, Japanese network, a network of the Koblenz Network Collection (KONECT) (Article–article wikilink). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 71,055,717 links among 1,610,638 nodes.

1.6M records1.6M nodes71.1M edges2013

Wikipedia Hyperlinks (Portuguese) Analyzed

Wikipedia links, Portuguese network, a network of the Koblenz Network Collection (KONECT) (Article–article wikilink). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 49,021,409 links among 1,603,222 nodes.

1.6M records1.6M nodes49M edges2013

Wikipedia Hyperlinks (Polish) Analyzed

Wikipedia links, Polish network, a network of the Koblenz Network Collection (KONECT) (Article–article wikilink). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 57,489,447 links among 1,529,135 nodes.

1.5M records1.5M nodes57.5M edges2013

Wikipedia French Talk Page Network Analyzed

Wikipedia talk, French network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 4,641,928 links among 1,420,367 nodes, timed from 20 June 2002 to 23 November 2015.

1.4M records1.4M nodes2.5M edges2016

Wikipedia Chinese Talk Page Network Analyzed

Wikipedia talk, Chinese network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 2,284,546 links among 1,219,241 nodes, timed from 17 November 2002 to 2 December 2015.

1.2M records1.2M nodes1.7M edges2016

Last.fm Friendship Network (December 2010, Syracuse Data Lab) Analyzed

The Last.fm friendship network of the Data Lab Social Computing Data Repository (Syracuse University): Last.fm, a music website founded in the United Kingdom in 2002. The network was crawled in December 2010 by Megan Kearl. This copy holds 5,115,300 directed links, each from a user to a friend of that user (the readme: the line 1,2 means user 1 is friend with user 2), among 1,191,812 distinct users, and a list of 108,493 user numbers (nodes.csv). Users are numbers; the files hold no names or profile data.

1.2M records1.2M nodes4.5M edges2010

Enron Email Network Analyzed

The Enron email network as KONECT distributes it: nodes are individual Enron employees and each row is one email between two of them, with the time it was sent. This copy holds 1,148,072 email links among 87,273 people, sent between 1 January 1980 and 26 May 2024.

1.1M records87.3K nodes321.9K edges2004

Wikipedia Arabic Talk Page Network Analyzed

Wikipedia talk, Arabic network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 1,913,103 links among 1,095,799 nodes, timed from 11 July 2003 to 23 November 2015.

1.1M records1.1M nodes1.6M edges2016

Wikipedia Italian Talk Page Network Analyzed

Wikipedia talk, Italian network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 3,067,680 links among 863,846 nodes, timed from 5 December 2002 to 23 November 2015.

863.8K records863.8K nodes1.7M edges2016

MySpace Aminer Social Graph (Follower Network) Analyzed

The social graph of MySpace, the social networking site with a strong emphasis on music, as distributed by AMiner and republished in Netzschleuder's file layout: 6,489,736 links among 854,498 users and a node table of 854,498 users with their IDs in the source data. Netzschleuder describes a directed edge (i, j) as user i following user j.

854.5K records854.5K nodes6.5M edges2013

Truth Social Dataset (Feb-Sep 2022) Analyzed

A crawl of the social network Truth Social made in 2022: 845,060 posts (called Truths; 244,206 are ReTruths and 283,191 are replies) by 17,377 authors with their like, ReTruth and reply counts, 454,458 user records (21,955 scraped profiles and 432,503 placeholders created for users seen only as authors or in follow links), 4,002,115 follow links, 506,276 reply links and 10,508 quote links between users, and the hashtags (21,599), media files (184,884), external URLs (173,947) and user tags attached to truths, collected by Patrick Gerard, Nicholas Botzer and Tim Weninger with the accompanying paper at ICWSM 2023 (doi 10.1609/icwsm.v17i1.22211). Time columns are text because the data state no time zone.

816.7K records454.5K nodes4.1M edges2023

Digg Friendship Network (December 2010) Analyzed

Digg is a social news website. This is the friendship network among its users, crawled in December 2010 by Mithila Nagendra, as distributed by the Data Lab's Social Computing Data Repository. This copy holds 7,261,524 friendship links among 771,231 users and a list of 116,893 user numbers.

771.2K records771.2K nodes7.3M edges2010

Wikisigned-k2 Signed Network Analyzed

A signed network of interactions between English Wikipedia users who edited pages about politics (the WikiSigned dataset of Maniu, Abdessalem and Cautis, based on 563 articles), in the form KONECT distributes it as wikisigned-k2. Interactions such as text editing, reverts, restores and votes were given a positive or negative value, so the network can be read as a web of trust. This copy holds 740,397 interactions among 138,592 users: 650,653 positive and 89,744 negative.

740.4K records138.6K nodes740.4K edges2011

Foursquare Friendship Network (December 2010) Analyzed

Foursquare is a location-based social networking website and mobile service. This is the friendship network among its users, crawled in December 2010 by Fred Morstatter, as distributed by the Data Lab's Social Computing Data Repository. This copy holds 3,473,834 friendship links among 639,014 users and a list of 106,218 user numbers.

639K records639K nodes3.5M edges2010

Wikipedia Portuguese Talk Page Network Analyzed

Wikipedia talk, Portuguese network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 2,424,962 links among 541,355 nodes, timed from 18 January 2004 to 23 November 2015.

541.4K records541.4K nodes1.5M edges2016

Wikipedia German Talk Page Network Analyzed

Wikipedia talk, German network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 6,729,794 links among 519,403 nodes, timed from 11 May 2002 to 2 December 2015.

519.4K records519.4K nodes1.8M edges2016

Wikipedia Spanish Talk Page Network Analyzed

Wikipedia talk, Spanish network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 2,702,879 links among 497,446 nodes, timed from 8 February 2002 to 3 December 2015.

497.4K records497.4K nodes1.3M edges2016

Twitter Social Network (Munmun) Analyzed

Twitter (ICWSM) network, a network of the Koblenz Network Collection (KONECT) (User–user following). KONECT's file for it has the header `% asym unweighted`, i.e. a directed network with weights marked `unweighted`. This copy holds 834,797 links among 465,017 nodes.

465K records465K nodes834.8K edges2010

Wikipedia Russian Talk Page Network Analyzed

Wikipedia talk, Russian network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 2,282,055 links among 457,017 nodes, timed from 26 December 2002 to 23 November 2015.

457K records457K nodes919.8K edges2016

Twitter Higgs Network (2012) Analyzed

Network Repository describes this dataset as one used to study the spreading processes on Twitter before, during and after the announcement of the discovery of a new particle with the features of the Higgs boson on 4 July 2012. The study (De Domenico and colleagues) covers messages posted from 1 to 7 July 2012. This file, soc-twitter-higgs, holds the links between Twitter accounts, which SNAP describes for the same study as a directed friends/followers graph; the retweet, reply and mention networks and the tweets are not part of it. This copy holds 14,855,875 links among 456,631 accounts.

456.6K records456.6K nodes14.9M edges2013

News Sharing Networks Expose Information Polluters on Social Media Analyzed

Twitter news-sharing networks built from COVID-related tweets to detect low-credibility accounts, as published by Truong, Allen and Menczer (OSoMe) on Zenodo: a retweet network among users (434,963 edges), a bipartite network of users and the news domains they shared (613,608 edges) and a user-user co-sharing network (48,770,798 edges), 49,819,369 edges in all. Edges carry a weight and the credibility labels and scores of their ends: a label of 0 (high), 1 (low) or -1 (unknown) and, for accounts, a score. Users are Twitter user IDs.

404.6K records404.6K nodes435K edges2024

Anatomy of an Online Misinformation Network (PLOS ONE 2018) Analyzed

Retweets of tweets that link to low-credibility ('claim') or fact-checking articles around the 2016 US presidential election, from the study 'Anatomy of an online misinformation network' (Shao and colleagues, PLOS ONE 2018), in the two files the archive provides: the file named pre-election has 2,132,235 retweets from 2016-05-16 to 2016-11-07, each labelled claim or fact-checking, and the file named post-election has 5,816,179 claim-only retweets from 2016-05-16 to 2017-10-09. The second file also contains all 1,746,867 claim rows of the first, so the two layers overlap; 7,948,414 rows in all. Each row is one retweet, with the user retweeted, the user retweeting, the tweet identifier and the time.

346.6K records346.6K nodes1.1M edges2018

Wikipedia Dutch Talk Page Network Analyzed

Wikipedia talk, Dutch network, a network of the Koblenz Network Collection (KONECT) (User–user messages). KONECT's file for it has the header `% asym positive`, i.e. a directed network with weights marked `positive`. This copy holds 1,554,699 links among 225,749 nodes, timed from 17 October 2002 to 23 November 2015.

225.7K records225.7K nodes565.5K edges2016

Libimseti.cz Dating Network Analyzed

Libimseti.cz network, a network of the Koblenz Network Collection (KONECT) (User–user profile rating). KONECT's file for it has the header `% asym weighted`, i.e. a directed network with weights marked `weighted`. This copy holds 17,359,346 links among 220,970 nodes.

221K records221K nodes17.4M edges2012

Google+ Social Network Analyzed

Google+ was a social network run by Google in which a member could add any other member to his circles, which creates a directed social graph. This is the Google+ network as Network Repository distributes it, as soc-google-plus, in a two-column edge list; the creators of the network are the authors of the paper that Network Repository cites for it. This copy holds 1,506,896 links among 211,187 members.

211.2K records211.2K nodes1.5M edges2013

Academia.edu Social Network Analyzed

Academia.edu is a platform on which academics share research papers. This is the network of links between its members as Network Repository distributes it, as soc-academia, in a two-column edge list; the creators credited here are the authors of the paper that Network Repository cites for it. This copy holds 1,398,063 links among 200,169 members.

200.2K records200.2K nodes1.4M edges2011

StackExchange SuperUser Interaction Network Analyzed

Super User is a question-and-answer site on the Stack Exchange network. SNAP builds temporal networks from the Stack Exchange data dump in which each interaction is a directed link (u, v, t): user u answered user v's question (a2q), commented on user v's question (c2q) or commented on user v's answer (c2a) at time t. Node numbers are the `OwnerUserId` of the dump, and t is a Unix timestamp in seconds. This copy holds the three interaction networks of Super User together, one row per interaction with its type: 430,033 answers to questions, 479,067 comments on questions and 534,239 comments on answers, 1,443,339 interactions among 194,085 users, posted between 2 August 2008 and 6 March 2016.

194.1K records194.1K nodes924.9K edges2017

Wikipedia Requests for Adminship Signed Network Analyzed

Votes in English Wikipedia's requests for adminship (RfA), the public elections in which editors support, oppose or stay neutral about a candidate for administrator, as SNAP publishes them with the voters' comments. Each row is one vote from a voter to a candidate: 198,275 votes among 11,381 users (144,451 support, 41,176 oppose, 12,648 neutral), cast between 1 May 2004 and 5 June 2013.

189.5K records11.4K nodes188.1K edges2014

Huawei Telecom Heterogeneous Graph Analyzed

An anonymised heterogeneous telecom graph provided by the Global Technology Services department of Huawei Technologies: 170,380 nodes of four types (1,000 app, 69,000 cell, 380 package, 100,000 user) with feature vectors, and 51,538,461 edges of three types (3,076,923 user-buy-package, 35,076,923 user-live-cell, 13,384,615 user-use-app) with attribute vectors. According to the README it represents the relationships between users and behaviours in a telecom network; node and edge attributes are floats normalised to [0, 1].

170.4K records170.4K nodes2M edges2021

Douban Friendship Network (December 2010) Analyzed

Douban.com is a Chinese web 2.0 site for reviews and recommendations of movies, books and music. This is the friendship network among its users, crawled in December 2010 by Long Qiu, as distributed by the Data Lab's Social Computing Data Repository. This copy holds 654,188 friendship links among 154,908 users and a list of 154,907 user numbers.

154.9K records154.9K nodes327.2K edges2010