truth_social — Truth Social users and interactions (2022)

Description

Networks of the users of Truth Social, the microblogging platform launched in February 2022 by the Trump Media & Technology Group, from the dataset of Gerard, Botzer and Weninger, who crawled the platform from February to October 2022 (posts of February to September), starting from popular accounts and following their followers and followings. The posts of Truth Social are called "Truths", and the reposts "ReTruths". The dataset contains 454,458 users, 4,002,115 follow relations and about 860,000 Truths with their replies, quotes, hashtags, user tags and links.

The entry contains seven networks:

follows: a directed edge from user u to user v means that u follows v.
replies: a directed edge from a user to a user they replied to, for each reply (the replied Truths are not recorded).
quotes: a directed edge from a user to the user whose Truth they quoted, for each quote.
retruths: a directed edge from a user to the author of a Truth they ReTruthed, for each ReTruth.
mentions: a directed edge from the author of a Truth to each user tagged in it, for each tag.
hashtags: a bipartite network of users and hashtags, with an edge for each use of a hashtag in a Truth of the user.
urls: a bipartite network of users and external URLs (links to other sites), with an edge for each link in a Truth of the user.

Repeated interactions appear as multiple edges, and self-interactions as self-loops. Each network contains only the users (and hashtags and URLs) with at least one edge.

The dataset was published in two versions on Zenodo, which differ only in the file of Truths: the second version leaves out 20,975 Truths whose texts contain quotation marks (malformed rows of the first version), and corrects the ReTruthed Truth of one. The Truths of both versions are combined here (those of the second version, and the 20,975 Truths only in the first one), so that the networks quotes, retruths, mentions, hashtags and urls include all of them. Even so, some Truths referenced by the files of tags, hashtags and links, and the original Truths of some ReTruths, are not in the file of Truths, so that their authors are unknown; these edges are left out (about 26,000 hashtag uses, 24,000 user tags, 22,000 links and 7,000 ReTruths). Rows of the file of Truths that are malformed (texts containing tabs) are read from the end, by the URL of the Truth. Also left out are the texts of the Truths, and the media files (images and videos) attached to them. The users and Truths that were added by the authors to complete the references have their unknown attributes set to -1. The times are given without time zone, and are taken as UTC; the creation dates of the users are known only to the month, and those of some Truths only to the day.

The networks have the following properties:

(Node) user_id (users): the id of the user in the dataset.
(Node) has_record (users): whether the user is in the file of users (a few users referenced by the follows, replies, quotes, tags and Truths are missing from it; their attributes are empty or -1).
(Node) username (users): the username.
(Node) created (users): the creation month of the account (Unix time; -1 if unknown).
(Node) follower_count, following_count (users): the numbers of followers and followed accounts shown on the profile (-1 if unknown).
(Node) time_scraped (users): the time when the profile was last scraped (Unix time).
(Node) finished_follower_scrape, finished_following_scrape, finished_truth_scrape (users): whether the followers, followings and Truths of the user were scraped completely at least once (1 or 0; -1 if unknown).
(Node) is_hashtag (hashtags): whether the node is a hashtag (otherwise a user).
(Node) hashtag (hashtags): the hashtag (empty for users).
(Node) first_seen (hashtags): the time when the hashtag was first seen (Unix time; -1 for users).
(Node) is_url (urls): whether the node is an external URL (otherwise a user).
(Node) url (urls): the URL (empty for users).
(Node) A few hashtags and URLs referenced by the Truths are missing from their files; their attributes are empty or -1. The user properties are empty or -1 for hashtags and URLs.
(Edge) time_scraped (follows, replies, quotes and mentions): the time when the edge was collected (Unix time).
(Edge) time (quotes, retruths, mentions, hashtags and urls): the time of the (quoting, ReTruthing, tagging, linking) Truth (Unix time; -1 if unknown).
(Edge) truth_id (quotes, retruths, mentions, hashtags and urls): the id of that Truth on Truth Social (its page is https://truthsocial.com/@<username>/posts/<truth_id>; -1 if the Truth is not in the file of Truths).
(Edge) quoted_truth_id (quotes): the id of the quoted Truth.
(Edge) retruthed_truth_id (retruths): the id of the ReTruthed Truth (-1 if unknown).
(Edge) like_count, retruth_count, reply_count (retruths, mentions, hashtags and urls): the numbers of likes, ReTruths and replies of that Truth when it was scraped (for ReTruths, as recorded, usually 0).

Tags

Social Online Political Bipartite Unweighted Multigraph Timestamps Metadata

Added

2026-10-31

Citation

P. Gerard, N. Botzer, T. Weninger, "Truth Social Dataset." Proceedings of the International AAAI Conference on Web and Social Media 17, 1034-1040 (2023) 10.1609/icwsm.v17i1.22211 Sci-Hub

Upstream URL OK

https://doi.org/10.5281/zenodo.7531625

Upstream license

Creative Commons Attribution 4.0 International (CC BY 4.0), https://creativecommons.org/licenses/by/4.0/

Networks

Tip: click on a column header to sort the list, or hover over it for a description.

Column legend

$\left<k\right>$: Mean degree
$\sigma_k$: Standard deviation of the degrees
$\lambda_h$: Largest eigenvalue of the Hashimoto (non-backtracking) matrix
$\tau$: Random walk mixing time
$r$: Degree assortativity
$c$: Global clustering coefficient
$\oslash$: (Pseudo-) diameter
$S$: Relative size of the largest component
NPs: Node properties
EPs: Edge properties

Name	Nodes	Edges	$\left<k\right>$	$\sigma_k$	$\lambda_h$	$\tau$	$r$	$c$	$\oslash$	$S$	Kind	Mode	NPs	EPs	gt	GraphML	GML	csv
follows	444,951	4,002,115	8.99	234.53	731.21	32.58	-0.44	0.01	6	1.00	Directed	Unipartite	user_id has_record username created follower_count following_count time_scraped finished_follower_scrape finished_following_scrape finished_truth_scrape	time_scraped	22.8 MiB	36.8 MiB	34.8 MiB	35.0 MiB
replies	56,974	506,276	8.89	189.55	160.21	21.62	-0.00	0.04	8	1.00	Directed	Unipartite	user_id has_record username created follower_count following_count time_scraped finished_follower_scrape finished_following_scrape finished_truth_scrape	time_scraped	2.8 MiB	4.6 MiB	4.5 MiB	4.2 MiB
quotes	3,092	10,508	3.40	36.61	28.91	22.17	-0.05	0.04	8	0.98	Directed	Unipartite	user_id has_record username created follower_count following_count time_scraped finished_follower_scrape finished_following_scrape finished_truth_scrape	time_scraped time truth_id quoted_truth_id	296 KiB	358 KiB	353 KiB	381 KiB
retruths	17,118	244,166	14.26	351.74	144.05	47.26	-0.10	0.05	8	1.00	Directed	Unipartite	user_id has_record username created follower_count following_count time_scraped finished_follower_scrape finished_following_scrape finished_truth_scrape	retruthed_truth_id time truth_id like_count retruth_count reply_count	3.3 MiB	4.4 MiB	4.2 MiB	5.4 MiB
mentions	17,986	121,553	6.76	177.58	52.81	680.77	-0.18	0.02	9	0.99	Directed	Unipartite	user_id has_record username created follower_count following_count time_scraped finished_follower_scrape finished_following_scrape finished_truth_scrape	time_scraped time truth_id like_count retruth_count reply_count	1.4 MiB	1.9 MiB	1.9 MiB	2.0 MiB
hashtags	25,428	187,584	14.75	131.73	106.54	35.92	-0.05	0.00	10	0.99	Undirected	Bipartite	user_id has_record username created follower_count following_count time_scraped finished_follower_scrape finished_following_scrape finished_truth_scrape is_hashtag hashtag first_seen	time truth_id like_count retruth_count reply_count	1.6 MiB	2.3 MiB	2.3 MiB	2.2 MiB
urls	163,820	231,086	2.82	43.71	58.00	8687.44	-0.24	0.00	14	0.97	Undirected	Bipartite	user_id has_record username created follower_count following_count time_scraped finished_follower_scrape finished_following_scrape finished_truth_scrape is_url url	time truth_id like_count retruth_count reply_count	7.8 MiB	10.5 MiB	10.2 MiB	10.8 MiB

Ridiculograms^*

^* These are automatically generated force-directed visualizations, and can be quite meaningless for networks both big and small. They should not be taken seriously as sources of scientific insight. See here for a discussion.