Inspiration

In the face of a situation as extreme as a pandemic, society also tends towards extreme behaviour. Nervousness and social hysteria due to confinement are present in our daily lives, and we see it reflected in the press, radio and television. As a consequence of this social hysteria, we constantly see cases of people breaking confinement, as well as doing massive purchases at the supermarkets. One example of this can be seen in the next link: Plunder in Italy supermarkets.

What it does

The idea we propose is to create a system to detect areas where social hysteria is increasing (based on the analysis and geolocation of posts in social networks), in order to detect those areas where riots are more likely to occur. This system would work as an early warning system, with the aim of alerting the authorities or the supermarkets themselves before the events occur, so that they can be warned of such behaviour.

How we built it

The first task we have is to study and understand how the narrative of social hysteria is developed. This task is vital because once we know how the narrative develops, we can focus on finding posts that follow this guidelines.

Secondly, we create a system to geolocate tweets in order to obtain the areas in which hysteria is increasing. We also develop a scrapper for newspaper headlines containing words related with Coronavirus and Social Hysteria. Last but not least, we gather info with Facebook and Google Trend APIs.

Finally, we aggregate the information gathered in a shiny app and the app periodically updates the state of hysteria in Spain and divided by its provinces.

Challenges we ran into

  • Generate a taxonomy of hysteric narrative in Spanish.
  • Obtain geolocated tweets and classify them according to their region.
  • Obtain global trends by Google and Facebook APIs.
  • Obtain information from newspapers.
  • Aggregate all the information gathered.

Accomplishments that we're proud of

We manage to combine several information resources such as newspapers, social networks and trends in order to detect possible riots based on social hysteria. This approach is crucial for the comming times. We are starting to register horrible news of the crowds calling for rebelion. With our approach we are making a call of duty to keep our relatives safe. Your knowledges do not make you great but what you do with them.

What we learned

By building the Social Hysteria detector we learned a lot of sociology in order to the understanding of the crowds and how they react to news and social events. Apart from that, we managed to organize ourselves as a team so that each member focused on what he does best. Another acquired knowledge was to aim all the information gathered towards something useful that served to support decision making processes.

Share this project:

Updates