During my research internship at Carnegie Mellon University, I investigated disinformation on Twitter using the BEND Framework, a set of 16 defined maneuvers that characterize influence campaigns on social networks. This project involves BEND labeling, a process in which datasets of 100 tweets from controversial topics were annotated with BEND maneuvers. The statistics presented with this program intends to present the inter-rater reliability in BEND annotations using Cohen's kappa in Python.
Further information regarding the program can be found in the presentation attached.