With increasing attention being paid to climate change every year, a recent study has been completed that applied data mining techniques, often applied to business or other big data problems, in order to discern patterns and gain insight into greenhouse gas emission patterns.
Dr Mohamed Gaber, from Robert Gordon University (RGU), Aberdeen, UK and Staci Lattimer, a postgraduate student at Newcastle University, UK, have applied a clustering algorithm to a dataset covering 21 years of environmental data from the UK and two countries of comparable population size - France and Italy.
‘With a vast amount of raw data being produced and collected, the environment sector is an area where the application of data mining could potentially have a huge impact. The need for understandable and useable information from this sector has increased, mainly thanks to awareness of issues such as climate change and carbon emissions,’ said Dr Gaber.
Topical issues such as climate change are often governed by politicians who do not necessarily have a technical background in science. Projects like this are key to breaking down the barriers of huge data sets to discern patterns that can not only shed light on the effects of climate change but also inform policy makers of important areas of high emissions where the greatest reductions can be achieved.
Analysing the data of individual industrial sectors, as well as the overall emissions, allows high and low emitting areas to be identified and appropriate decisions taken to reduce emissions.
The study showed some interesting results: for 18 of the 21 years analysed, the UK’s greenhouse gas emissions were overall much lower than France and Italy. However, French emissions showed a decreasing trend over the period, whereas the UK’s increased.
The worst polluter for each of the three countries was the energy sector, while solvents and other product use showed the lowest emissions.
The UK leads the others in the waste industry, which is ranked third highest emitter in the UK and second highest in both France and Italy. The UK managed to reduce waste incineration emissions, where France and Italy did not, with the UK's waste emissions being around 20% lower than both France and Italy.
‘While statistical analysis is often used in relation to greenhouse gas emissions, data mining could provide a more efficient and powerful way of looking for unknown trends,’ said Gaber. He went on to say: ‘It is hoped that the results from this project demonstrate the power of data mining by identifying hidden patterns in order to encourage more analysis to be done in the area of climate change, as there are many other areas to be explored.’
‘Making the results easier to understand and interpret by the end user could lead to more meaningful actions by the authorities, especially in topical issues such as climate change and energy management, as often the people who make the decisions on data mining are not of a technical data mining background.