Go back

Real-time social research will offer fresh insights

New tools for analysing online media will let researchers and the public watch social processes as they unfold. The project shows how social scientists can respond to the rise of corporate Big Data, say Rob Procter, William Housley and Adam Edwards.

In their quest to spot emerging trends, corporations have developed infrastructure and strategies to collect many types of data from their customers. They use such information to group people into important populations that are inaccessible to public social science or, indeed, any meaningful public scrutiny.

The rise of corporate Big Data has fed into the debate among social scientists about the broader purpose and status of social research. In 2005, Michael Burawoy, the former president of the American Sociological Association, urged academics to create a “public sociology” capable of sustaining, nurturing and defending civil society against state and market pressures.

Two years later, Mike Savage and Roger Burrows wrote a paper arguing that the emergence of big and broad data streams generated by commercial organisations was compromising the capacity of social science to take on that role. The ever-growing volume of data collected by businesses via loyalty schemes, credit cards and e-commerce has fuelled the fear that social scientists’ research methods are becoming obsolete.

But as well as helping to create this problem, online technologies can provide some solutions. The use of social networking sites, increasingly accessed through mobile devices, to access information and network with friends and like-minded strangers is creating data that are highly significant for social research. At the same time, the increasing adoption of open-data principles by large public bodies is allowing researchers to interrogate more streams of digital data.

In 2014, researchers will be able to use data to study unfolding social processes at the population level via Cosmos, the Collaborative Online Social Media Observatory. Funded by the Economic and Social Research Council and the UK’s education technology charity Jisc, Cosmos harvests freely available data from blogs, RSS feeds and microblogs such as Twitter, as well as openly available data sets from, for example, the Office for National Statistics.

Cosmos provides a suite of tools for analysing text and networks. From a single tweet, for example, it can automatically extract information about the writer’s gender and location, and the tweet’s sentiment and tension scores. At the aggregate level, Cosmos provides tools to visualise tweeters’ social networks and use of keywords and hashtags, and to detect patterns in retweeting, topic and location. These individual tools can be combined to form an analytical workflow for more complex analysis. The first version of Cosmos is undergoing beta testing, and additional tools are in development.

The project has grown out of a collaboration by sociologists and computer scientists at Cardiff University and the Universities of Warwick, Edinburgh and St Andrews, aimed at developing ways to harvest and analyse social data, and to create tools that can detect tension and cohesion in online social networks. An initial example involved the analysis of social-media streams connected to football and alleged racist name-calling by players.

One Cosmos project analyses the origin and spread of hateful speech and antagonistic content. An example of this is the study of homophobic content directed at high-profile sportspeople. Another project is investigating the value of social-media data in building accurate statistical models for predicting crime in major cities.

Cosmos is a resource for public sociology with citizen participation at its heart. We aim to make our research findings quickly and easily accessible, and are exploring ways to promote ‘citizen social science’, whereby members of the public can get involved in the research. Volunteers will be able to access social research collected by Cosmos and, by performing simple annotation tasks, help to improve the quality of computer-based analyses.

This area is changing fast, and we’re still not sure how emerging techniques will affect research. It may be that these forms of data will replace more traditional methods such as sample surveys and in-depth interviews. Or social research may reorientate around new objects, populations and techniques. But the best outcome would be for emerging methods to be used alongside existing ones, to make research richer and more nuanced.

Social science is not the only discipline in which knowledge production is becoming more public, but it’s probably the one in which the change seems most appropriate. Research can now be carried out and communicated differently, forcing us to rethink the role of the academic social scientist.

Rob Procter is professor of social informatics at the University of Warwick; William Housley is professor of sociology at Cardiff University; and Adam Edwards is a senior lecturer in sociology at Cardiff University and director of the Cardiff Centre for Crime, Law and Justice. All are members of the Cosmos project team at www.cosmosproject.net

More to say? Email comment@ResearchResearch.com