IMI project studying fake news exposure in Switzerland (in French, German, and Italian).
In this project, we focus on the networks to understand how fake news are spreaded. We study user networks and filter bubbles (the effect when users are trapped in a community (like in a bubble), which prevents them from being exposed to alternative opinions).
You can click on the links below to explore the network view of the data we extracted from social media.
Here, we extract networks of Twitter accounts spreading information about controvercial topics.
- General controvercial topics
- Conspiracy
- Hierarchy. Communities discussing controvercial topics are less hierarchical.
Here, we study various Youtube networks. We have two lists of videos. One contains videos about controvercial topics. We call them sources of information. The other has videos that were identified as fake by an expert.
First, we extract networks of videos, users, and channels. Two videos are connected if Youtube recommends watching one video after another. Two channels are connected if they feature each other or share a video. Two users are connected if they interact in the comment section.
List of pre-selected seed videos
List of pre-selected seed videos
These videos are not very popular, so very few users comment on them. Therefore, the network is very small and doesn't bring any interesting insights. We don't extract users network for this topic.
List of pre-selected seed videos
*Youtube data is collected using Youtube Data Tools by Bernhard Rieder