Characterizing Regional and Behavioral Device Variations Across the Twitter Timeline

Identifying Sentiment Influences Provoked by Context Factors – Results from a Data Analytics Procedure Performed on Tweets

“…However, the authors do not match these influences with continents (cf. [42]) as we did within our paper. To address this research gap, we performed analysis that combine mobile devices' OS and continents.…”

“…Tweets that are sent from mobile devices tend to be more negative in terms of expressed sentiments than those that are sent from desktops [40,41]. As for mobile devices' OS, more positive posts are more likely to stem from blackberry devices [42].…”

“…[20])) instead of whole continents. [42] analyzed the influence of mobile OS on sentiments. However, the authors do not match these influences with continents (cf.…”

“…Tweets are amongst others characterized by their shortness (e.g. [31,36,42,44,45]). Since November 2017, the maximum number of characters to be used within a tweet is set to 280 (cf.…”

“…Furthermore, our analysis has to cope with the huge number of users around the world that post tweets (cf. [31,39,42,44,48]). Therefore, we assume that the methods for processing the tweets need to be fast and performant and must deal with the shortness of tweets to obtain accurate and reliable results [45].…”