Abstract:
|
CSIRO have developed the "we Feel" web based analytics tool that monitors emotions (love, joy, surprise, sadness, fear, anger and others) for the Black Dog Institute. This monitoring is world-wide and for this paper the data are split into the following regions North America (USA and Canada roughly 60%), South America (7.6%), Europe (16%), Asia (11%), Africa (2.9%), Oceania (1.7%). The aim of the study is to detect unusual changes in emotions over time. The data runs from 1 January 2015 to the end of 2017. The We Feel system is built for real-time analysis of the emotional sentiment expressed on Twitter. The large majority of Twitter accounts are publicly available, allowing users to view posted tweets. This contrasts with other social network platforms which have different privacy settings. The default data source for the We Feel system is a custom stream created using the Twitter query interface to return all English language tweets containing a custom vocabulary of emotional words. In addition to the emotional categorisation, the We Feel system classifies tweets based on the location of the user. For more detail see Larsen et al. (2014) and Paris et al (2015).
|