Building off of Datayo’s topic modeling toolkit, we break down North Korea’s official English news media by sentiment. First we explore how sentiment patterns manifest within the entire English language news corpus from 1996 to 2020. Then we look at how sentiment patterns manifest within a set of articles related to our previously identified nuclear weapons topic.
We find that articles discussing nuclear weapons are more negative on average compared to the rest of the corpus. We also find that nuclear documents with negative sentiment utilize words that highlight the threat of foreign actors towards North Korea’s current government. Alongside this, we find that nuclear documents with positive sentiment utilize words more associated with domestic pride and development.