Big, Open Data: Mining and Synthesis (BODMAS)

The volume and assortment of available data for research in the social sciences has dramatically increased in recent years- a trend that shows no sign of stopping. For the first time researchers can obtain large amounts of population data free of charge (so-called "open data") thanks to government websites such as data.gov.uk. When these data are combined with the computing power to perform complex calculations it creates an unprecedented opportunity for social science researchers. We are now in an era of big data and this is fundamentally changing the research environment for investigations across social science. The purpose of this project is to develop some of the new perspectives required to adapt to these changes in the practice of data modeling and synthesis.

These new perspectives include the need to account for the increased uncertainty in data provenance and less thorough metadata, as the data provision philosophy has shifted away from careful collection and dissemination to an emphasis on expediency. Researchers increasingly have to temper gains in data volume against losses in data quality when they embark on a study. Extra caution is also required when combining datasets, especially if they contain geographic information, as it is not always case that the spatial scales are compatible. The proposed project will develop a web-based tool to help social scientists minimise or eradicate these issues by enabling the synthesis, mining and visualisation of open datasets in a more informed way. The project will also use the newly combined data to undertake more complex analyses of population processes using supercomputers to gain unprecedented insights into phenomena such as commuter flows.

In addition the project is focused facilitating my personal ambition to become a Future Research Leader. A comprehensive list of world-leading collaborators (ESRI (UK), the Open Data Institute, University of Illinois and University of Zurich) each have a specialism of interest to me and that is integral to the project. Activities with these organisations and my mentor, Professor Michael Batty, form part of a comprehensive plan for knowledge transfer and personal development. I have the full support of my host department, the UCL Centre for Advanced Spatial Analysis, which ensures that the project activities are not confined to the time formally costed to it. As the proposal demonstrates, the proposed project is both ambitious and extremely timely and will strive for high impact social science.

1.1 THE ESRC
Influencing Behaviour and Informing Interventions: The project is about linkage and synthesis across a range of datasets from administrative, commercial and volunteered sources. The challenge associated with this is nontrivial and if successfully overcome it will have important impacts on the interaction of data between seemingly disparate sources. Additionally, the range of international partners selected in this project reflect expertise that will be important to UK social sciences. It is hoped that the outcomes from this project can help to inform future ESRC data strategy through the incorporation and development of the ADT findings at the start of the project through to the integration of datasets with the ESRC Data Service at the end.
A Vibrant and Fair Society: The largest impact this project will have is the better use of data in the social sciences. By utilising new technologies, a number of high-impact data visualisations are envisaged which, will raise wider public awareness of society's inequalities. Wider outreach is detailed in the proposal, but much of it will be motivated by the development of a vibrant and fair society.
1.2 ACADEMIC PARTNERS (UNIVERSITY OF ZURICH AND UNIVERSITY OF ILLINOIS)
I have expressed an interest in presenting my research at the departmental seminars at each of the universities I intend to visit to inspire further collaborations with world leading researchers. The ambition to co-author peer reviewed publications will also provide a lasting outcome from the relationship that, it is hoped will extend beyond the duration of the project. In addition collaborators will be able to utilise some of the carefully prepared datasets assembled thus expanding the applications and relevance of their own research. My work with the University of Illinois will, for example, help to build a compelling case for greater involvement of the social sciences in US cyberinfrastructure projects and this, in turn, will contribute to the case for continued funding of the host department, whilst my work with Zurich could link to a broader proposal to the European Science Foundation.
1.3 INDUSTRY
If investment in open data, especially from government sources, is to be sustained and justified there need to be tangible benefits beyond academic research. Through partnering with ESRI (UK) this project will help inform the company about both the possible value of open data (something they already have an awareness of) and best practice to unlock its potential. In addition I have agreed a series of industry workshops with the Open Data Institute (ODI) to showcase the data and tools created and compiled over the course of this project.
1.4 SOCIAL SCIENCE RESEARCH
The ambition to publish not only provides a quantifiable method of demonstrating research output, it will also offer focus to the range of collaborations I plan to pursue. Work with the University of Illinois will yield a paper on the value of supercomputing to social science, with the University of Zurich a paper on using visual analytics to gauge the quality of open data in geography and with my academic mentor a wide ranging paper on flows in population data. I will aim for highly cited international and interdisciplinary journals and believe that the strength of the work I plan to undertake, combined with the interest in the topic, will facilitate this.
1.5 POLICY
The proposed policy document will go some way to ensuring the promises of open data are realised by offering clear guidance and examples of best practice. The document itself is not meant to replace or override existing larger scale initiatives (such as the ADT); it merely seeks to contribute to the growing body of literature from the perspective of GIScience to better inform constructive access to open data.