Pandemic Social Media Hate Speech Analysis

A group of experts (hardly known to the public before the pandemic) all of a sudden had to deal with social media interactions.

Using insius InTrack Technologie we could retrieve using the Twitter Stream API

6.2 mio. tweets which results in “only” 3.6 mio. tweets when excluding retweets
The data comes in live and unfiltered and still contains reported and deleted tweets before their actual deletion.

Several phenomenen are being discovered:

  • One impulse from an expert results in hundreds of responses or more
  • The possibilities of replying to someone’s opinion have resulted in rants of backfiring at the initial professional input
  • On the long term experts are silenced due to this very unusual situation

Our goal is to better understand the mechanisms which may turn a message into a hate message and provide numbers about the situation in 2022.

On the technical side we try to better understand the opportunities of discovering hate speech in distinction from offensive speech with Artificial Intelligence Tools.

Presentation Holger Lütters and André Lang during #GOR 2022 at HTW Berlin


Learn More about insius InTrack:

Follow Holger Lütters on Researchgate