russian_troll_mentions — Russian troll Twitter mentions (2012-2018)

Description

Directed network of mentions among Twitter accounts linked to the Internet Research Agency (IRA), a Russian "troll factory". The tweets were collected by Darren Linvill and Patrick Warren (Clemson University), who gathered, using the Social Studio tool, nearly 3 million tweets sent between February 2012 and May 2018 (mostly from 2015 to 2017) from the IRA-linked handles that Twitter identified and provided to the US Congress in November 2017 and June 2018, and coded the accounts by their theme. The tweets were published by FiveThirtyEight. The network was constructed by Christopher S. Marcum from the version of the tweets published on August 1, 2018: tweet authors and contents were lower-cased, the handles mentioned with "@" were extracted from the tweet texts (which include retweets), and only mentions of handles that are themselves authors in the data were kept. Each node is a troll account (only accounts that mention or are mentioned by another troll account are included), and a directed edge from account i to account j indicates that i mentioned j in at least one tweet. Self-mentions were discarded when the network was built, so that 33 accounts (presumably those that mentioned only themselves) are isolated nodes.

The network has the following properties:

Nodes: name, the Twitter handle (lower case); activity, the number of tweets by the account in the data; account_type, the specific theme of the account, as coded by Linvill and Warren (e.g. "Right", "left", "Russian", "Hashtager", "Koch"); account_category, the general theme of the account, as coded by Linvill and Warren ("RightTroll", "LeftTroll", "NewsFeed", "HashtagGamer", "Fearmonger", "NonEnglish", "Commercial", "Unknown"); region, the list of region classifications of the account's tweets, as determined by Social Studio (an empty string means a tweet without region); language, the list of languages of the account's tweets; max_following, min_following, the maximum and minimum number of accounts followed by the account at the time of its tweets; max_followers, min_followers, the maximum and minimum number of followers of the account at the time of its tweets; first_post, last_post, the dates and times of the first and last tweets of the account (as "YYYY-MM-DD HH:MM:SS").
Edges: mentions, the number of mentions of the target in tweets by the source (repeated mentions in the same tweet are counted separately).

Tags

Social Online Weighted Metadata

Citation

D. L. Linvill, P. L. Warren, "Troll Factories: Manufacturing Specialized Disinformation on Twitter." Political Communication 37(4), 447-467 (2020) 10.1080/10584609.2020.1718257 Sci-Hub
C. S. Marcum, "Russian Troll Twitter Mention Network." GitHub repository (2018) github.com
O. Roeder, "Why We're Sharing 3 Million Russian Troll Tweets." FiveThirtyEight (2018) fivethirtyeight.com

Upstream URL OK

https://github.com/cmarcum/RussianTrollNetwork

Networks

Tip: hover over a column header 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
russian_troll_mentions	1,245	2,974	2.39	14.19	17.87	315.59	-0.34	0.06	15	0.92	Directed	Unipartite	name activity account_type account_category region language max_following min_following max_followers min_followers first_post last_post	mentions	38 KiB	55 KiB	54 KiB	50 KiB