Lawrence Ampofo and Ben O'Loughlin will present a paper on social media monitoring at the CRESC Annual Conference on 31 August - 3 September 2010, University of Oxford. Previous CRESC conferences have been an excellent forum bringing together social theory and methodology and this year's theme, 'The Social Life of Methods' allows us to explore the intriguing possibility that our methods dictate our knowledge and action rather than being tools we use to shape social change.
Our paper, 'Real-Time Social Media Monitoring: Automated Mass Observation?', presents a new method of mining and extracting insight from social media feeds. It exemplifies the dilemma of how new forms of tracing and visualising social interactions may channel the questions we ask and the conclusions we seek to draw. Designed by experts in computational linguistics at Linguamatics and researchers at Royal Holloway's New Political Communication Unit, the methodology Real-Time Social Media Monitoring allows researchers to aggregate and analyse data from social media platforms as an event or crisis unfolds to inform timely decision making. The tools and techniques that constitute the methodology can be used for a broad range of purposes in politics and business, such as identifying shifting brand reputations, key opinion leaders, viral content, and emergent groupings, networks and the geolocation of citizens/users. The paper demonstrates this through a series of case studies examining public responses to H1N1 vaccine take-up, the 2009 Copenhagen climate summit, Haiti's earthquake, and the 2010 UK General Election.
The presentation of this methodology and the case studies raise a number of methodological, conceptual and ethical questions. Methodologically, how do we visualise social networks as they evolve in real-time and how do our visualisations feed back into policy interventions? How can we maintain validity of conceptualisation, measurement and inference in social media analysis? And how do we discern intent? For instance, acts of expression-for-itself, persuasion, and deception all occur in our cases. Conceptually, what does our methodology imply for traditional distinctions of public/private, broadcast/dialogue and directed/emergent communications? Ethically, is social media monitoring simply a non-intrusive instant polling technique, or a form of Mass Observation for the 21st Century?