The DEMICOM research centre at Mid-Sweden University just released their fast-tracked popular science report "Stjärnspäckat" ['Star Spangled'], including 70 contributions about the US Election by researchers and writers from a range of perspectives.
DIGSUM director Simon Lindgren's contribution is based on an analysis of bots on Twitter during the run-up to the election. The results are summarised in a network visualisation (green dots = humans, black dots = bots). The main conclusions are that:
- 1/3 of the most active accounts are bots.
- 1/3 of the most central (in terms of network metrics) accounts are bots.
- Bots are everywhere in the network. In no key clusters are bots absent.
- In terms of volume and centrality, Biden's and Trump's accounts are relatively marginal on the Twitter platform as such. Even though their tweets are of course picked up by news media, and spread in other ways.
The fact that politics is becoming increasingly automated and algorithmically driven poses challenges for researchers, citizens, politicians, as well as for democracy.