The «fake news» term has really taken all its meaning during the harsh 2017 presidential campaign. Today, it has become the plague of society, if most of the time it seems obvious, it becomes increasingly difficult to recognize it. But whose fault it is?
Social networks are so much a part of our lives, that false information can spread all over the planet. Our will to share everything fast is against our critical mind and verification of information. Fake news can reach millions of people who have full confidence in the people they follow and the number of likes.
During today’s global health crisis many questions are raised but few answers are given: fake news has exploded.
That is why we decided to work on CovidoScope to offer Internet users tools to become aware of the misinformation around covid-19. For this purpose, we use the analysis of true and fake tweets.
With CovidoScope, we want to provide the user and the wider world with the so-called «silent» characteristics of tweets
dealing with the Covid-19 epidemic.
As Rumor Gauge research is complex, we decided to focus on features that are easy to obtain but offer the most meaningful information possible.
The goal is not to create a tool to determine if a tweet is true
or not, but rather to offer a galaxy of tweets about the pandemic and leave it to the user to analyze the impact of fake news versus lambdas tweets.
Thus, we decided to use the «Covid-19 Misinformation» dataset by Shahan Ali Memon and Kathleen M. Carley, offering a selection of 4574 Tweets posted between January and August 2020 by people living in the United States. Each post was analyzed and classified in no less than 17 categories (e.g. fake news, politics, fake treatment, true prevention, etc).
The existence of these categories opens different possibilities of exploitation.
We then decided to represent the number of retweets per category as a function of time with its geographical parameters and its statistics, which would allow us to have a first overview of the temporal dynamic.