What is computational social science?

What is Computational Social Science?

The amount of data tracked by online apps and social media platforms has skyrocketed in recent years. This means researchers and analysts now have access to an almost infinite pool of data on the social behaviour of people online. Computational social science is a field in which researchers use this data to answer important social science questions. To do this well, they need to develop data analysis skills and programming expertise.

Why does it matter?

When people leave a digital footprint recorded in the form of text, images, interaction metrics, and of course – metadata, the motivations driving those posts can be analysed and processed mathematically to produce estimates related to human social behaviour. Researchers work to combine existing big data sets from places like Twitter, comments on blog posts, or key words in the media, with small-scale randomised survey responses to better understand human culture in the digital age. These systems then help researchers answer questions that are important to policy makers, NGOs, businesses, and society more broadly.

Poverty

The applications of computational social science are vast and can offer insights into global issues such as poverty. The UN estimates that the global poverty rate is projected to be seven percent in 2030. The ongoing COVID-19 crisis, violent conflict and climate change all pose real threats to reducing global poverty. Another key challenge in the fight against poverty is a lack of good data. Many developing countries don’t have national statistical systems that provide information about where and to what degree people are suffering from poverty.

To fill this gap, researchers working in computational social science have found ways to provide estimates of levels of wealth tied to residential locations through the analysis of mobile phone metadata. Drawing upon existing metadata from the call logs of 1.5million mobile phone users in Rwanda, researchers were able to analyse the locations of incoming and outgoing calls in connection to a randomised survey of 1000 mobile phone users to build a model that can estimate levels of poverty in residential locations.

The estimates developed by computational social science researchers can be used by policy makers and NGOs to direct aid and development strategies towards targeted areas showing the highest levels of poverty. By combining questions related to social justice with big data, this research is scalable and can be applied to a range of different contexts and social settings.

Climate change

In the summer of 2019 and 2020, Australia experienced months of catastrophic bushfires. With 33 human deaths, nearly 3 billion animals killed or displaced, and over 24 million hectares burned, the Australian bushfires attracted significant public attention and much debate on social media. During the fires and in their immediate aftermath, many scientists, politicians, and climate activists were hopeful that the scale of destruction would lead to action on climate change.

In response to this crisis, researchers working in computational social science sought to better understand the relationship between the bushfires and perceptions around climate change. Drawing on a dataset of 9000 tweets, researchers were able to identify and track key words and hashtags to assess perceptions of blame, causality, levels of emergency and prevention tactics related to the bushfire crisis. In doing so, they were able to analyse a highly charged and emotional debate through scientific means.  

They found that despite the spread of some disinformation, Twitter activity around the Australian bushfires led to a strengthening of support for action on climate change. Analysis showed that users engaging in bushfire related debates on Twitter sought to delegitimise those who were perceived to be responsible for the fires, such as climate denying politicians. Not only was it possible to track an increase in discussions relating the fires to climate change, but the nature of these discussions was used to draw greater attention to the issue itself.

What can we expect for the future of CSS?

The field of computational social science is growing and now includes researchers from across the globe. Driven by the goal to reduce barriers to access for early career researchers, two scholars – sociologist Chris Bail (Duke) and an information scholar Matt Salganik (Princeton) have launched the Summer Institutes in Computational Social Science (SICSS). SICSS is an international state-of-the-art training program created to train the next generation of researchers and incubate cutting-edge projects that transcend disciplinary boundaries.

In June 2022, SICSS has been held in Australia for the first time. It is hosted at the Sydney Social Sciences and Humanities Advanced Research Centre (SSSHARC) and co-sponsored by CSIRO Data61 and the Academy of the Social Sciences in Australia. It opened its doors to an inaugural cohort of PhD students and early career researchers from 10 Australian universities working across 12 disciplines, from media and communications, to mathematics, to law. By bringing together researchers and industry partners from the technology sector, SICSS-Sydney aims to support the next generation of computational science experts working to address pressing global issues.