At the beginning was Big Data, then Machine Learning came and now is the time of AI. Since buzzwords emerge faster than the innovation waves they describe, the conceptualization of digital transformation remains sometimes vague, although it can be considered the result of convergence among advances made in several related information and communication technologies (ICTs). “Data is the new oil” claimed the marketing guru Clive Humby back in 2006 and since then this mantra has been obsessively relaunched by academics, practitioners and policymakers. Although most of them are quite confident about the disruptive effect of data on our economies and societies, it seems we are just at the beginning of a paradigmatic change, yet uncertain of its direction and consequences.
The combination of internet, mobile communication and the Internet of Things (IoT) has kept on generating an enormous amount of information and, as information processing capacity become available, digital information technology induces structural changes in individual human behaviour and in the collective (social) behaviour of human societies.
The data revolution has dramatically improved the computational and explanatory capacity of many hard sciences, but social sciences have also been deeply affected. Digital transformation has changed both how produce and consume and has informed almost every choice we make, whether we rent a car, buy a pair of shoes, catch a bus, book a hotel room, apply for a job or date a complete stranger.
Yet, less attention has been paid to the work of those social scientists who are supposed to use this information to shed new light on old problems and understand the undergoing transformation of life in our communities. In other words, to fully appreciate the transformative power of research over our society, we should try to understand how the data revolution has transformed our disciplines and our job as social scientists.
What are the challenges that the science of data has offered economists, sociologists, political scientists, psychologists, geographers and policymakers, just to mention a few? How can this enhanced data paradigm generate meaningful insights for external stakeholders and society as a whole? What are the risks of this orientation to quantify and measure behaviours?
We’ll try to address these and similar questions during the panel “Data for good. How Data has transformed today’s Social Sciences?” being held on the 16th September 2019 as part of the University of Birmingham’s University Research Conference (#BhamResCon19).
The transdisciplinary panel (h.14.25 at the Great Hall) will be coordinated by Dr Massimiliano Nuccio and will include talks by Prof. Raquel Ortega-Argiles (City REDI – UoB), Prof. Siddhartha Bandyopadhyay (Centre for Crime, Justice and Policing – UoB) and Rebecca Riley (West Midlands Combined Authority and City REDI).
Data transformation is probably making economics a less “dismal science” and, hopefully, will help to provide us with more effective economic recipes. We can use data to classify, to explain, and to forecast. On the one hand, data has expanded the potential for economic modelling and improved our understanding of firms, markets and consumers. On the other hand, the economists’ toolbox has opened to new societal challenges and exchanges with other related disciplines like history, sociology, psychology.
The diffusion of urban analytics has transformed many cities into labs for innovation. Spatial and real-time data are dramatically changing mobility and transportation and provide citizens with opportunities to become smarter consumers, opening up opportunities for innovation in the framework of the circular and sharing economy.
In the field of political studies, data on individual behaviours have contributed to innovative and critical approaches to the analysis of “crime”, “justice” and “policing”. At the central and local government level, we are on the way to test data-driven policies. Big data and advanced analytics can potentially improve the design and monitoring of public policies and eventually shall reduce the gap between politics and citizens for a more shared and informed decision-making.
From the perspective of the philosophy of science, the challenge of data science to traditional social sciences is paramount and could deeply affect the foundations of academic research as we know it. Big data analytics can push to a resurgence of inductive over deductive research aimed at discovering a hidden or latent pattern in data before formulating any theoretical hypothesis. Shall we fear a possible return to empiricism and to “the end of theory” as it has been provocatively labelled in the famous article by Chris Anderson (2008)? The combination of algorithm and data visualization unleashes a huge potential to classify features and objects, to build complex networks and to predict different and apparently unrelated social and economic phenomena. Nevertheless, social sciences are well aware that correlation and co-occurrence do not imply causation and, although a good narrative of a phenomenon is pivotal to its understanding, they cannot rely on only perceptions and descriptions.
We tend to agree with those like Rob Kitchin (2014) who calls instead for defining a new paradigm of “data-driven science” which could, for instance, combine prediction and causation in a coherent research framework. According to this author, social sciences and humanities can only benefit from leaving behind neo-positivistic approaches which have been the ground for ill-advised policies and attain more breadth, depth and reliability in current understanding of human behaviour.
This blog was written by Massimiliano Nuccio, Research Fellow, City-REDI.
The views expressed in this analysis post are those of the authors and not necessarily those of City-REDI or the University of Birmingham.
To sign up to our blog mailing list, please click here.