This paper aims to present a methodology for monitoring and assessing the public perception of security on social media, given the great importance of security in determining the people’s well-being. The methodology is applied to different corpora obtained by collecting Twitter messages about three topics of the agenda setting related to security. It operates in two steps, in the first one, ETM is performed in order to identify topic representation and sentiment. In the second step, this information is transformed in a stream of numerical data and a composite index aggregates the information on the sentiment related to the three topics. The composite index embeds a penalty function, which reduces the weight of the sequences, which show the greatest volatility over time. Results show that this procedure allows for real-time measurement of the perception of security, which is mostly negative. It is net of the effect of communication hype, and it enables to quantify reliably any potential change in the public perception.
