Twitter is adding those accounts and their tweets to a public database it launched previous year to track its battle against government-linked misinformation.
Social media giant Twitter on Thursday released a new archive of state-backed propaganda from accounts it has banned based in Iran, Russia, Spain, and Venezuela.
The removed accounts and their posts are being included in the company's public archive that tracks its efforts against government-linked misinformation. Twitter said the removals are meant to prevent election interference while preserving valid political speech. The remaining 2,865 accounts "employed a range of false personas to target conversations about political and social issues in Iran and globally". Anyone can download the datasets, which has some information redacted. Now the data set contains more than 30 million tweets and more than one terabyte of images and video.
Most of the Iranian accounts were found to be spreading news stories angled to support Iranian geopolitical interests or to be fake user profiles created to manipulate online debate.
A smaller sub-group, originating in Iran, exclusively "engaged with discussions related to Israel".
In addition to information about Iranian accounts, Twitter's most recent numbers showed it removed four accounts from Russia as part of its investigation into the Russian Internet Research Agency (IRA), a center of state actors in St Petersburg that sent out more than 10m tweets attempting to influence USA politics between 2013 and 2018.
The St. Petersburg-based "troll farm" has been accused by US intelligence agencies of working with Russian intelligence to sow discord and spread misinformation in the run-up to the 2016 USA presidential election.
"Our further analysis suggests that they were operated by a commercial entity originating in Venezuela", the post said.
The U.S. platform said it had also taken down 130 accounts tied to the Catalan independence movement in Spain.
Mr Roth said Twitter had collected all the suspect accounts it had removed since beginning its purge into a data set that it provided to researchers.