Statistics of co-occurring keywords in confined text messages on Twitter
1 Niels Bohr Institute, University of Copenhagen, Blegdamsvej 17, 2100 Copenhagen, Denmark
2 Physics of Geological Processes, Department of Physics, University of Oslo, PO 1048, Blindern, 0316 Oslo, Norway
Received: 30 June 2014
Revised: 21 July 2014
Published online: 22 September 2014
Online social media such as the micro-blogging site Twitter has become a rich source of real-time data on online human behaviors. Here we analyze the occurrence and co-occurrence frequency of keywords in user posts on Twitter. From the occurrence rate of major international brand names, we provide examples of predictions of brand-user behaviors. From the co-occurrence rates, we further analyze the user-perceived relationships between international brand names and construct the corresponding relationship networks. In general the user activity on Twitter is highly intermittent and we show that the occurrence rate of brand names forms a highly correlated time signal.
© EDP Sciences, Springer-Verlag, 2014