Applying data science

Answering society’s most pressing questions

In an increasingly data-driven age, governments, organisations and researchers are questioning how to harness the good that can come from responsible use of data while at the same time minimising inherent risks to individuals, groups, and society at large.

By leveraging its position as a world leader in social science research, LSE aims to develop the necessary tools for analysing large data sets and to answer questions regarding how organisations can make well informed data-driven decisions.

A Data Science Initiative at LSE will support interdisciplinary data science research and teaching across the School’s academic disciplines. It will form a hub for researchers to collaborate on data-driven research to address major challenges to governments, businesses and society.

Some examples of the use of data science are below.

Better health outcomes
Data science has the potential to transform the healthcare sector, whether through defining and accessing relevant data, or increasing patient engagement through digital solutions. LSE Health is working with China’s National Center for Cardiovascular Diseases (NCCD) on leveraging data sets and health surveys to evaluate the effectiveness of treatment methods and settings, ultimately aiming to better support governmental action in addressing major burdens of disease. LSE Health is also leading a pan-European initiative that will provide methodological guidance to a range of disease-specific projects. Working with public and private partners, LSE will lead on common practices in data collection and usage, working to set the agenda for future investment in this area.

Understanding uncertainty in environmental modelling
Modelling and simulation are an increasingly important part of modern science, especially in highly policy-relevant disciplines such as weather, climate and hydrology. Good practice in the use and interpretation of models is therefore vital, both for sound science and for informing evidence-based policy decisions. LSE’s Centre for the Analysis of Time Series (CATS) sits at the heart of this approach to data analysis. Researchers in CATS have done groundbreaking research on climate science, energy futures, mathematical chaos, medical signals analysis, and weather forecasting, among other things. CATS has active partnerships with EDF Energy, Lloyd’s of London, the National Centre for Atmospheric Research, the Royal National Lifeboat Institution, and The Start Fund.

Systemic risk and massive official data sets
The global financial crisis has renewed academics’ and policymakers’ interest in understanding the nature of the danger posed by the financial system to society, in particular how financial risk is endogenously generated in the marketplace and can be hidden from authorities until too late. Yet research is stymied by the practical challenges of analysing the massive available data on the activities of financial institutions. The Department of Finance and the Systemic Risk Centre (SRC) are working with a number of central banks, supervisors and private sector companies to help them overcome this technical barrier; using data – from public to highly confidential – to increase understanding of financial risk, to clarify the network of interconnections between financial actors and to formulate more robust macroprudential policies.

How cities perform Global urban growth is increasing the challenge of governing cities. At the same time, there is little data on how cities are governed and what capacity they have to address complex and interrelated economic, social and environmental challenges. LSE Cities is analysing and mapping official data sources to reveal cities’ social, governance, planning, transport and environmental patterns, determining whether they have the necessary capacity or if new arrangements are required to ensure cities are able to play a fundamental role in reducing global energy demand and limiting carbon emissions, as well as responding to issues of social inclusion. The Centre, through the Urban Age Programme, further investigates how the physical and social are interconnected in cities, comparing those in rapidly urbanising regions in Africa and Asia, as well as in mature urban regions in the Americas and Europe.

Planned activities

Our future growth is limited only by resources. With sufficient support, we would like to:

  • Locate in a purpose-build physical space, with facilities for hosting staff, students and visitors, and for computing, collaboration, hosting workshops, and visualization;
  • Provide more support for the Director through partial buyout;
  • Appoint a Centre Manager;
  • Appoint a second Data Science Fellow to staff a new programme in LSE Data Science Surgeries and proto-type Incubator model. The innovative data science incubator model pioneered by the University of Washington and New York University’s Centre for Data Sciences enables cross-disciplinary progress by bringing together data scientists and domain scientists to work on focused, intensive, collaborative projects, frequently involving a non-trivial software engineering or methodological component presenting practical obstacles for social scientist researchers. The incubator model offers researchers from any field the chance to learn through collaboration with data science experts how to surmount these obstacles, not by outsourcing these tasks but by helping the domain experts actively learn the technical solutions through the project incubation process. The incubator programme will consist of two dedicated data science fellows hosted at the LSE but drawn from the DSI, available full time to assist with the challenges of large-scale data processing, linking data from different sources, and using high-performance computing resources of ICL, including big data processing and analysis. A third social science research fellow with data science experience will be hosted at the DSI, to engage data scientists based at Imperial College;
  • Operate an inter-disciplinary seminar series, drawn from speakers from Imperial, LSE, and externally;
  • Host a set of workshops for researchers, to learn new techniques in data science and computation;
  • Host visiting scholars and those on sabbatical, from partner institutions, other universities, or industry;
  • Issue “challenge calls” for innovative solutions, including “hackathon” events to solve specific problems or contests to come up with innovative solutions for targeted questions – an innovative solution-oriented approach called for in the recent FinTech Futures report;
  • Organise data policy “trials”, for prototyping innovative policy solutions and working with civil society organizations and entrepreneurs involved in innovative data practice;
  • Promote outreach and education. A key activity of the Centre will be outreach to practitioners and industry partners, forming a core part of its rationale in contributing knowledge to policy in the data economy and information society. This outreach will maximize the impact of the Centre’s research, with a multiplier effect of spreading awareness of social big data and data science issues through education, seminars in methodology and data management, continuing education and executive education, along with potential revenue-generating activities for sustainability and fostering partnerships with industry and government. We have two new MSc programmes commencing in 2017 and another one from 2018. See the Study page for further information.

Current priorities:

  • The Department of Statistics is pushing the boundaries in developing new statistical and algorithmic methods in the areas of machine learning, statistical learning for complex and dependent data, social statistics and risk management, and complex time series analysis. Specific areas of research include change point detection, functional regression, computation-statistical trade-offs, and scalable algorithms for processing streaming data, as well as data with a combinatorial structure such as social graph data.
  • The Department of Mathematics is developing new solutions in the areas of combinatorial optimisation, financial mathematics, game theory and operations research. Academics have expertise in: analysing the sample complexity of machine learning, addressing how much data is required for reliable conclusions; the development of novel machine learning methods, including those based on various notions of ‘definitive’ classification; and analysis and development of fast algorithms – including sub-linear algorithms for tasks such as property testing.
  • The Department of Methodology focuses on the application of social research methods, which include innovative qualitative approaches to novel data, rather than being restricted to just applied statistics and computational methods. The department hosts world renowned experts who research and teach on topics as diverse as measurement problems, event forecasting, computational text analysis, the structure and design of causal inference, and experimental and quasi-experimental methods for digital ethnography and surveying. Its world-leading, methodologically innovative research covers applications such as social media, election analysis, criminology, political competition, health policy, and urban studies.
Twitter Auto Publish Powered By :