Chain-sense signature of the Twitter interaction network during the 2016−2022 period

Abstract

Measuring the structure of interaction networks through time allows for studying when and how changes in the behavior of a social network platform and its users occur. Here, we use chain-sense, which can be roughly described as a way to describe the linearity of interactions, to measure and describe the interaction network sampled from the Twitter public stream API from 2016 to 2022. We compare the measurements before and after 2020, which marked the start of lockdowns due to COVID-19 and the increase in internet activity. The chain-sense signature post-2020 indicate a bimodality different from those observed prior.