Building daily economic sentiment indicators

The availability of copious amounts of data produced by the increasing datification of our society is nowadays deemed an opportunity to produce timely and convenient statistical information. This paper shows the building of economic sentiment indexes from the texts of the most read economic newspapers in Spain. The data are collected through the scraping of the Digital Periodical and Newspaper Library website. To compute the sentiment, an existing emotional lexicon for Spanish words has been customized, allowing for inferring sentiment for words in texts. The resulting indexes are later compared to other well-known indicators that try to monitor similar or related phenomena.