Benchmarking Micro-Blog Performance: Twitter Content Classification Framework

Benchmarking Micro-Blog Performance: Twitter Content Classification Framework

Stephen Dann (Australian National University, Australia)
DOI: 10.4018/978-1-4666-8408-9.ch014


This chapter outlines a content classification framework designed to categorize content from individual and group Twitter activity. Measurement of Twitter at the individual account level can support the analysis of individual use of Twitter, and, guide the use of the platform for commercial operations. Applying a pre-existing content classification framework allows for the consistent coding of Twitter timelines into one of the five categories, with an option to further refine into a series of sub-categories. This coding approach allows for the ongoing longitudinal measurement, benchmark and analysis of how individuals or groups use their social media accounts. This chapter also outlines the potential use of the classification framework as a planning tool for guiding content creation. This approach creates a two-stage process of planned content engagement and consistent content measurement metrics from a single framework.
Chapter Preview


Twitter launched in 2006 as a short messaging service intended to replicate and extend the limited text message capacity of mobile phones to provide a group-wide communication platform. It has become a mainstay of the text based micro-blogging and life streaming platforms as it provides a combination of real-time broadcast and delayed response. As the platform has evolved, it has moved through varying emphasis on the web presence to a combination of mobile and desktop based activity. In early 2014, Twitter redesigned the timeline experience to reinforce the emphasis on the service as a web-based experience. Although less than a decade old, Twitter has become a distinctive media channel to supplement television and live events, as a host medium for ad hoc community and personal conversations. Prior research on Twitter divides into broad categories of macro-data sets, legitimization studies and behavioral measurement.

Macro-data papers cluster around the use of Twitter as a sentiment collection engine with an emphasis on machine based coding (Dodds, Harris, Kloumann, Bliss, & Danforth, 2011). High volumes of Twitter are used for predictive insights (Bollen, Mao, & Zheng, 2011) to uncover a social barometer of public sentiment (Bollen, Pepe, & Mao, 2009; Thelwall, Buckley, & Paltoglou, 2011). Legitimization papers focus on the validation of micro-blogging. These papers study Twitter in a range of contexts for political activity (Cetina, 2009), civil unrest (Fahmi, 2009) and social activism (Galer-Unti, 2009). Specific roles for Twitter in journalism (Gay, Plait, Raddick, Cain, & Lakdawalla, 2009) for eyewitness accounts (Lariscy, Avery, Sweetser, & Howes, 2009) or natural disaster detection (Longueville, Smith, & Luraschi, 2009) emerged from this research field. Behavioral Measurement examines emergent behaviors such as the use of the @ symbol marking replies (Naaman, Boase, & Lai, 2010) #hashtags (boyd, Golder, & Lotan, 2010), and the “retweet” protocol (Java, Song, Finin, & Tseng, 2007). It also explores Twitter’s role for personal proximity (Hohl, 2009) connectedness (Henneburg, Scammell, & O'Shaughnessy, 2009) conversation (Steiner, 2009) and friendships (Fernando, 2010).

Complete Chapter List

Search this Book: