CaSMa – Citizen centric approaches to social media analysis

The spread of the Internet into nearly every aspect of our daily lives has stimulated the growth of a vibrant and expanding set of systems to support social interaction via a range of internet-connected devices. Often collectively referred to as social media this set of applications and services are undergoing a highly unstable and dynamic growth as society shapes, and is shaped by, the emergence of these new capabilities.  The sheer volume of users of these systems and the diversity of people using social media make it immediately appealing as a means of gaining insight and understanding into human behaviour.

People have turned to social media for a broad range of purposes. They have expressed their views on critical issues that support their personal relationships; they have expressed support and comfort to others; they have coordinated political actions and even responded to major crises[i]. In doing so users of social media have started to express many facets of human behaviour and have revealed many aspects of their identity – including their political beliefs, sexuality, spirituality – and offered information about how this might relate to others[ii]. Consequently, researchers have turned to social media as a critical source of information about human behaviour and social interaction. Two key drivers have dominated these investigations:

• Understanding the ways people use social media and what this means for people and society
• Understanding social phenomena and events expressed in social media by drawing upon social media as a critical source of information.

To gain insight and understanding researchers have exploited a variety of Data Analytic techniques. These have been used to uncover key personality traits[iii], predict and even influence societal outcomes[iv] , and to provide accounts of critical societal events (e.g. the London Riots[v], the Arab Spring[vi]).  The resulting combination of widespread data capture, data mining, machine learning and large scale cloud computing is often collectively referred to as “big data” and has already being hailed as a major revolution in understanding human behaviour[vii].

The information available in dominant social media services such as Facebook and Twitter represents only a small fraction of the personal information we generate. From Web searches and browsing histories, purchases, texts, emails, photos, videos, locations, levels of activity, heart rates and stress levels – the list of human data we bleed as we go about our everyday life continues to grow and is increasingly collected and used by both commercial and government organizations. Indeed it has increasingly become recognised as a major business and societal asset[viii].

However, there is also a growing public concern over the exploitation of personal data, with recent press articles claiming that companies can uncover highly personal events such as pregnancy[ix] or the likelihood of divorce[x] before we even know ourselves, raising concerns about the analysis of personal communications by the intelligence services[xi] or questioning what happens to our data when we die[xii]. In turn, organisations as diverse as Liberty[xiii] and Google[xiv] have called for greater transparency, trust and security in the use and analysis of our personal data.

Over the last five years Horizon has explored the capture and use of a growing contextual footprint and how this human data might be used to gain critical insights into various aspects of peoples lives. Our fundamental objective is to reap the benefits offered by the growing raft of social media and personal information while avoiding the potential pitfalls that might arise from the analysis of human data. Two challenges are critical to horizon’s orientation to human data

• The development of facilities and approaches that are sensitive to the personal nature of human data
• The promotion of responsible innovation in the capture, analysis and use of human data.

Within this project we wish to build upon our current approaches to develop a set of services and facilities for the social sciences. These will allow researchers to draw upon social media and associated human data to gain insights into human behaviour in a manner that is ethically sound.

The CaSMa project was supported by ESRC grant ES/M00161X/1

[i]       See http://www.ifrc.org/PageFiles/134658/WDR%202013%20complete.pdf