Uncover stories in COVID-19 data
I used AI as an investigative tool to reveal hidden stories in social media data. I processed millions of Twitter data to surface the topics that were mostly associated with the word Covid and the different sentiments on the pandemic expressed by the media as opposed to Italians.
I worked on this project with the aim of showcasing the power that AI can provide in journalism by revealing hidden patterns in data. By formulating some questions that I wanted to investigate, I found the hypothesis to test: understanding whether or not there was a difference in how media outlines talked about Covid versus how people on Twitter talked about it. What topics were associated with the word Covid? What was the sentiment of the people towards the virus versus the sentiment of the news headlines published by the media? These questions led to my research into the data, and that led to the writing of the story.
Data-driven story for Esquire Magazine
I first processed ten months of Twitter's data, which accounts for millions of tweets. By developing an NLP algorithm and using Topic Modeling techniques within R studio, I managed to portray the digital conversations of the Italian news media outlets and the conversations occurring on Twitter over the two lockdowns that hit Italy between February 2020 and November 2020. I then visualized my insights in support of my storyline and wrote a data-driven story published on Esquire in December 2020.
The collaboration with Esquire offered me the opportunity to build a story upon AI, by utilizing Natural Langue Processing (NLP) techniques.
To discover and build the story I first processed 10 months of Twitter's data, which accounts for millions of tweets. By developing an NLP algorithm and using Topic Modeling techniques, I managed to portray the digital conversations of the Italian news media outlets occurring on Twitter over the two lockdowns that hit Italy, starting from February 2020.