The truth between the lines
Insight
Fake news, rumors, unverified or expired content: misinformation imposes its social agenda by playing a decisive role in political debate at the local level. People post untrusted content for various reasons, and while in most cases it is good faith or superficiality, in other cases it is for propaganda purposes or worst, manipulation.
After considerating the reliability of the institutional (but not only) media at the local level, the Emilia Romagna Region has involved us in the analytical process of the impact of fake news on the territory.
Solution
We decided to adopt an approach based on information bubbles by relying on Advanced Analytics techniques and Machine Learning regression mechanisms. This method was mostly based on:
- the ease of applying behavioral analysis to Twitter, a social media widely used to disseminate information content;
- the possibility of training the model to recognize fake news from a relatively small information base;
- the universal validity of the model training phase.
The system designed provides daily analysis and supports the constant growth of the size of the dataset. Thanks to a scalable architecture, the precision and good performance in data processing is guaranteed.
Benefit
Emilia Romagna Region has managed to identify who has the habit of posting fake news inside a specific catchment area thanks to the aggregation of the results dashboard and which are the reference sources and volumes of fake news around.
