Use case: Coronavirus in Brazil – Sentiment Analysis of stories in the media on a government under critics

Use case: Coronavirus in Brazil – Sentiment Analysis of stories in the media on a government under critics

This use case focuses on the demonstration of some of the capabilities of SAIL LABS Media Mining System (MMS) illustrating the media coverage on the coronavirus pandemic outbreak.

To download the short use case as a pdf, please click on the button below.

  • Allows continuous assessment of media coverage and public opinion in real time.
  • Suitable for monitoring punctual and permanent health promotion campaigns.
  • Sentiment Analysis of traditional and social media regarding the perceived impact of a disease and government action.
  • Event-based surveillance.
  • Rapid data collection for fast decisions.
  • Timely reporting mechanisms.
  • Social media intelligence.
  • Enables to extract data from text or audiovisual sources in languages that the users do not speak thanks to automatic speech recognition and machine translation covering 32 languages.
  • Enables situational awareness and situational understanding.

In this example, the results provided by SAIL LABS Media Mining System help us to identify stories in the media related to the profile “Brazil” and “coronavirus”. The system enables us to discover results in Portuguese and to understand better the political instability that arises in the biggest country in South America during the coronavirus outbreak.
As the local press states, governors who declared states quarantined are being criticized by the president Jair Bolsonaro. Besides still reassuring Brazilians that the infection is “just a little flu” and the corona crisis a “hysteria”, the president called the governor of the state of Sao Paulo (44 million habitants) a “lunatic”, because he declared a 15 day quarantine (Figure 1). Regarding the attempt by the governor of Rio de Janeiro to close the city’s airport, the president said that this is a measure that only “generates a climate of terror in the population that can lead to depression and lower immunity”. Both states have the highest number of cases and fatalities (Source: Carta Capital, accessed via SAIL LABS Media Mining System).

Moreover, as soon as Brazilians heard about the possibility of cure trough “hydroxychloroquine”, the medication used in the treatment of other diseases, such as malaria, the drug got sold out in many cities (Figure 2).
Two days later, Jair Bolsonaro made announced that the medication is a possibility of cure for coronavirus and that the army would be mobilized to increase the production of the medication (Figure 3).
Additionally, militia and drug dealers in favelas in Rio de Janeiro have imposed a curfew in order to halt the spread of the new coronavirus (Figure 4). Messages were sent by the social networks, and indicate to the population to not leave home after 8:00 pm. “Attention all residents of Rio das Pedras, Muzema and Tijuquinha! Curfew from today 20:00. Whoever is seen on the street after this time will learn to respect the next one!”, “We want the best for the population. If the Government doesn’t have the capacity to fix it, organized crime will solve it”, indicates another message (Source: SIC Noticias, accessed via SAIL LABS MMS).

The MMSys and the Sentiment Analysis tool can be used to evaluate Jair Bolsonaro´s government approval reflected in the media. Considering that Brazil announced its first coronavirus case on the 26th of February, we can compare results from two time frames. From the 26th of February to the 8th of March, the system analyses the sentiment of 12931 stories in the media. The majority of stories associated with the profile “Jair Bolsonaro” (32.16%) were classified in the negative sentiment category (Figure 5). When analysing the period from the 09.03 to 26.03, a total of 40.5% stories were classified as having a negative sentiment (Figure 6). Figure 7 points out the volume trend of stories classified with positive and negative sentiment.

Therefore, the SAIL LABS Media Mining System Sentiment Analysis can be used to provide insights on the attitude of a speaker or a writer with respect to some topic or the overall contextual polarity of a document, serving as a guide to better understand the acceptance or rejection of a topic of interest. It can also be used combined with a trend graph, indicating the maximum amount of media stories per day, according to the sentiment selected.

Request the Complete Coronavirus Report