The open source air pollution project

The basic premise of this project is that the vast amount of air pollution data collected in the UK and worldwide is massively underutilised. However, innovative analysis techniques, and a platform through which to make them accessible, are now available for end-users to exploit the value of these data. This project will make an important contribution to unlocking the informatics value locked up in these data, with concomitant benefits to end-users. The careful and targeted analysis of air pollution data can yield many important insights concerning source contributions and the underlying chemical and physical processes, leading to improved decisions and a stronger science base relating to the management of air pollution. This project approaches the problem in an innovative way which aims to maximise the benefits to a wide range of users by making air pollution analysis tools available for free in an open-source environment. Air pollution remains an important topic worldwide, and in Europe is the environmental factor with the greatest impact on health and is responsible for the largest burden of environment-related disease, with estimated UK health costs of £9.1/21.4 billion in 2005 due to particles alone. The measurement of pollutant concentrations is fundamental to the effective management of air pollution, and as a result the number of continuous measurement sites continues to grow due to the efforts of central and local government, industry and academia. In the UK alone it is estimated that there are > 1000 continuous air pollution monitoring sites that collect inorganic gases and a range of particle metrics. The knowledge being transferred in this project relates mostly to the results from research funded over the past 3-5 years. Its principal application and contribution to improving quality of life relates to the wider access and adoption of innovative methods to critically analyse and understand air pollution data. These methods include graphical-statistical approaches for source identification and quantification, meteorological normalisation using statistical modelling to account for meteorological effects in time series, and trend and change-point analysis. This research is backed up by extensive publications by the applicant in leading peer-reviewed journals. In addition to making available these methods to the air pollution community, this project also aims to develop a framework for the analysis of air pollution data. Currently, no such framework exists and the analysis of air pollution data by end users is fragmented and of mixed quality leading to an inconsistent approach to analysis. A core component of this project is the development of tools in a software system called 'R'. R is a highly capable programming language that has been specifically designed for analysing data. Its core capabilities include leading data analysis, graphing and statistical techniques, which have been greatly extended through the more than 1200 packages developed by users worldwide. R is open-source and free and works on all common computing platforms, thus encouraging its very widespread and growing use. It has been identified as the ideal tool through which the aims of this project are delivered. The project will make available analysis tools primarily through the development of an R package, which will be made freely available through a global network of package repositories. Working with the open-source software R will help to ensure the continued sustainability of this project beyond its 36 month duration. The R community is a vibrant, growing and internationally distributed community of people, which themselves include academics, regulators, private sector employees and leading statistical computing scientists. This project approach is designed to encourage others to contribute to the tools, either by refining existing methods or implementing new ones.