The economy and the web: reti a regional economic twitter index

In recent years, data from social media are increasingly used to monitor opinions, forecast, and analyse economic and social dynamics. In this paper we develop RETI, a Regional Economic Twitter Index capable of analysing economic dynamics at the regional level, monitoring daily posts on the social media Twitter. Starting from April 2020, we automatically downloaded tweets referred to the North-East of Italy and containing economic terms, selected using state of the art word-embedding techniques. We applied Sentiment Analysis techniques, using an unsupervised lexincon-based approach and computed the economic sentiment of users. The index dynamics is coherent with real world facts and media information and is positively correlated with standard economic indicators computed by ISTAT, confirming the ability of social media of capturing facts and short run confidence in the economy.