Sentiment analysis on social network data: the Regional Index RETI

Data from social networks has been used more and more in recent years to track opinions, make predictions and study social and economic dynamics. In this study, we create the Regional Economic Twitter Index (RETI), which can analyse regional economic dynamics by keeping track of daily tweets on the social media Twitter. We automatically downloaded tweets that mention the North-East of Italy and contain economic terms, starting in April 2020. These tweets were chosen using cutting-edge word-embedding techniques. Utilizing an unsupervised lexicon-based approach, we applied sentiment analysis techniques and determined the users’ economic sentiment. The index dynamics are consistent with news reports and information from the real world, and it exhibits a positive and statistically significant correlation with ISTAT’s standard economic indicators, demonstrating social media’s capacity to capture information quickly and accurately.