15A - Arleth Guerrero
- Leveraging social media data during pandemics: A visual analytics approach
Nowadays, social media platforms generate an immense amount of information in the form of text, images, video, sound, and others. Their capabilities and reliability during adverse situations have made them society’s go-to communication method as they continue to operate while more traditional methods fail (Wallop, 2014). With the unexpected arrival of the COVID-19 pandemic, billions of tweets were generated, bringing both opportunity and challenges to emergency managers when seeking to leverage social media data as a source of information. Therefore, this research investigates how emergency managers could utilize social media data for monitoring the aggregated published tweets to enhance their strategic decision-making process. To do so, we have adapted a visual analytics framework that has been developed for monitoring public sentiment and extracting patterns of social media during product recalls (Zavala and Ramirez-Marquez, 2019). The proposed work understands that by developing an alert warning system based on collective sentiment analysis, decision makers will be able to identify and anticipate scenarios where significant levels of negative sentiment has been disseminated. We implement Data Analytics, Natural Language Processing, and Machine Learning techniques to generate inferences from microblogging Big Data. We extracted a sample of more than 60 million tweets, which were filtered to the city of El Paso, Texas to develop a proof-of-concept for the proposed alert warning system. Results indicate that, the adapted framework could assist emergency managers when seeking to monitor social media; however, it was found that additional bias challenges must be addressed before it is fully implemented.