How useful are social networks for analyzing epidemics? The example of Twitter for monitoring the 2018–2019 Ebola epidemic in Africa

The last major Ebola outbreak in Africa occurred in 2014-2015. This epidemic has mainly affected three West African countries (Guinea, Sierra Leone, Liberia). She killed 20, 000 people. A number of articles have studied the rumors that circulated during this outbreak on Twitter. For example, Fung's article pointed out that these media disseminated false information about the treatment of the disease, such as bathing in salt water to heal (Fung, 2016). Jin's article also pointed out that these media were behind a fake news of a snake at the origin of the epidemic. This article also listed the Top10 rumors circulating on twitter (Jin, 2014). Ebola virus disease has been raging in the Democratic Republic of Congo (DRC) since 1 August 2018. It killed more than 2050 people. It is the second largest epidemic after West Africa. No studies have been conducted to determine who is communicating about this epidemic and what types of tweets/rumours are being disseminated? To answer this question, we conducted an analysis on Twitter via Radarly
^®
software, over a period from April 1 to July 7, 2019. The keyword Ebola and #Ebola were used, with a filter on the French language. 17282 tweets were collected and classified via the software. The tool also extracted and represented the knowledge in a map-like way (volume of publications / time...). It also identified the dominant themes in the form of clusters. The tone of the messages has also been determined. Our work has highlighted several actors communicating around the epidemic: the general public, experts, politicians, the media... Concerning the general public, it was pointed out that the Congolese population communicating on the social network was shared. Some actors accepted the disease and acted as relays of preventive messages and fight against false rumors. But another party considered the disease to be a kind of conspiracy to destabilize the country and denied it. This part of the population spread false rumours and accused health workers of spraying the virus. Concerning the experts, our study revealed that they shared many tweets to inform the general public. Concerning the media/press, they are both Congolese and international. In particular, it was pointed out that they were disseminating prevention messages, including response teams. Concerning the politicians, they supported medical teams by relaying educational messages. Our study found 12 main influencers and revealed a negative message tone of 73.31 %. However, our study suffers from bias. We did not take into account other media such as Facebook and we limited our study to the French language.