Log in
Since 2008, statistical research at the University of Bristol has significantly influenced policies, practices and tools aimed at evaluating and promoting the quality of institutional and student learning in the education sector in the UK and internationally. These developments have also spread beyond the education sector and influence the inferential methods employed across government and other sectors. The underpinning research develops methodologies and a much-used suite of associated software packages that allows effective inference from complicated data structures, which are not well-modelled using traditional statistical techniques that assume homogeneity across observational units. The ability to analyse complicated data (such as pupil performance measures when measured alongside school, classroom, context and community factors) has resulted in a significant transformation of government and institutional policies and their practices in the UK, and recommendations in Organisation for Economic Co-operation and Development (OECD) policy documents. These techniques for transforming complex data into useful evidence are well-used across the UK civil service, with consequent policy shifts in areas such as higher education admissions and the REF2014 equality and diversity criteria.
Small area estimation (SAE) describes the use of Bayesian modelling of survey and administrative data in order to provide estimates of survey responses at a much finer level than is possible from the survey alone. Over the recent past, academic publications have mostly targeted the development of the methodology for SAE using small-scale examples. Only predictions on the basis of realistically sized samples have the potential to impact on governance and our contribution is to fill a niche by delivering such SAEs on a national scale through the use of a scaling method. The impact case study concerns the use of these small area predictions to develop disease-level predictions for some 8,000 GPs in England and so to produce a funding formula for use in primary care that has informed the allocation of billions of pounds of NHS money. The value of the model has been recognised in NHS guidelines. The methodology has begun to have impact in other areas, including the BIS `Skills for Life' survey.
Visual analytics is a powerful method for understanding large and complex datasets that makes information accessible to non-statistically trained users. The Non-linearity and Complexity Research Group (NCRG) developed several fundamental algorithms and brought them to users by developing interactive software tools (e.g. Netlab pattern analysis toolbox in 2002 (more than 40,000 downloads), Data Visualisation and Modelling System (DVMS) in 2012).
Industrial products. These software tools are used by industrial partners (Pfizer, Dstl) in their business activities. The algorithms have been integrated into a commercial tool (p:IGI) used in geochemical analysis for oil and gas exploration with a 60% share of the worldwide market.
Improving business performance. As an enabling technology, visual analytics has played an important role in the data analysis that has led to the development of new products, such as the Body Volume Index, and the enhancement of existing products (Wheelright: automated vehicle tyre pressure measurement).
Impact on practitioners. The software is used to educate and train skilled people internationally in more than 6 different institutions and is also used by finance professionals.
Targeted Projection Pursuit (TPP) — developed at Northumbria University — is a novel method for interactive exploration of high-dimension data sets without loss of information. The TPP method performs better than current dimension-reduction methods since it finds projections that best approximate a target view enhanced by certain prior knowledge about the data. "Valley Care" provides a Telecare service to over 5,000 customers as part of Northumbria Healthcare NHS Foundation Trust, and delivers a core service for vulnerable and elderly people (receiving an estimated 129,000 calls per annum) that allows them to live independently and remain in their homes longer. The service informs a wider UK ageing community as part of the NHS Foundation Trust.
Applying our research enabled the managers of Valley Care to establish the volume, type and frequency of calls, identify users at high risk, and to inform the manufacturers of the equipment how to update the database software. This enabled Valley Care managers and staff to analyse the information quickly in order to plan efficiently the work of call operators and social care workers. Our study also provided knowledge about usage patterns of the technology and valuably identified clients at high risk of falls. This is the first time that mathematical and statistical analysis of data sets of this type has been done in the UK and Europe.
As a result of applying the TPP method to its Call Centre multivariate data, Valley Care has been able to transform the quality and efficiency of its service, while operating within the same budget.
This case study concerns the development and subsequent uptake of the Feature Selective Validation (FSV) method for data comparisons. The method has been adopted as the core of IEEE Standard 1597.1: a `first of its kind' standard on validation of computational electromagnetics and is seeing increasingly wide adoption in industry practice where comparison of data is needed, indicating the reach and significance of this work. The technique was developed by, and under the guidance of, Dr Alistair Duffy, who has remained the world-leading researcher in the field. The first paper on the subject was published in 1997 with key papers being published in 2006.
Wavelets and multiscale methods were introduced and rapidly became popular in scientific academic communities, particularly mathematical sciences, from the mid-1980s. Wavelets are important because they permit more realistic modelling of many real-world phenomena compared to previous techniques, as well as being fast and efficient. Bristol's research into wavelets started in 1993, has flourished and continues today. Multiscale methods are increasingly employed outside academia. Examples are given here of post-2008 impact in central banking, marketing, finance, R&D in manufacturing industry and commercial software, all originating from research at Bristol. Much of the impact has been generated from the original research via software. This software includes freeware, distributed via international online repositories, and major commercial software, such as Matlab (a preeminent numerical computing environment and programming language with over one million users worldwide).
This impact case is based on economic impact through improved forecasting technology. It shows how research in pattern recognition by Professor Henry Wu at the School of Electrical Engineering and Computer Science led to significantly improved accuracy of daily national gas demand forecasting by National Grid plc. The underpinning research on predicting non-linear time series began around 2002 and the resulting new prediction methodology is applied on a daily basis by National Grid plc since December 2011. The main beneficiaries from the improved accuracy (by 0.5 to 1 million cubic meters per day) are UK gas shippers, who by conservative estimates save approximately £3.5M per year. Savings made by gas shippers benefit the whole economy since they reduce the energy bills of end users.
Research conducted in UCL's Department of Statistical Science has led to the development of a state-of-the-art software package for generating synthetic weather sequences, which has been widely adopted, both in the UK and abroad. The synthetic sequences are used by engineers and policymakers when assessing the effectiveness of potential mitigation and management strategies for weather-related hazards such as floods. In the UK, the software package is used for engineering design; for example, to inform the design of flood defences. In Australia it is being used to inform climate change adaptation strategies. Another significant impact is that UCL's analysis of rainfall trends in southwest Western Australia directly supported the decision of the state's Department of Water to approve the expansion of a seawater desalination plant at a cost of around AUS$450 million. The capacity of the plant was doubled to 100 billion litres per year in January 2013 and it now produces nearly one third of Perth's water supply.
Reversible Jump Markov chain Monte Carlo, introduced by Peter Green [1] in 1995, was the first generic technique for conducting the computations necessary for joint Bayesian inference about models and their parameters, and it remains by far the most widely used, 18 years after its introduction. The paper has been (by September 2013) cited over 3800 times in the academic literature, according to Google Scholar, the vast majority of the citing articles being outside statistics and mathematics. This case study, however, focusses on substantive applications outside academic research altogether, in the geophysical sciences, ecology and the environment, agriculture, medicine, social science, commerce and engineering.
A generalized additive model (GAM) explores the extent to which a single output variable of a complex system in a noisy environment can be described by a sum of smooth functions of several input variables.
Bath research has substantially improved the estimation and formulation of GAMs and hence
This improved statistical infrastructure has resulted in improved data analysis by practitioners in fields such as natural resource management, energy load prediction, environmental impact assessment, climate policy, epidemiology, finance and economics. In REF impact terms, such changes in practice by practitioners leads ultimately to direct economic and societal benefits, health benefits and policy changes. Below, these impacts are illustrated via two specific examples: (1) use of the methods by the energy company EDF for electricity load forecasting and (2) their use in environmental management. The statistical methods are implemented in R via the software package mgcv, largely written at Bath. As a `recommended' R package mgcv has also contributed to the global growth of R, which currently has an estimated 1.2M business users worldwide [A].