ABSTRACT
Introduction In Chapters 3 and 4 we have shown how social media can be profitable when used to derive insights on electoral campaigns, to nowcast the campaign dynamics and to forecast the final results. While most of the attempts presented in those chapters were successful, we also observed some variation in the accuracy of the predictions. This is true not only for our analyses (consider for instance the selection of the PD party leader: see Chapter 3) but also for the analyses made by other scholars who have employed alternative techniques (see Chapter 1).
