Handbook

There has been an explosion in the availability, scope and amount of data that is available for organisations to analyse in order to help them make better decisions. Big Data has presented new challenges for data analysts and has meant that traditional statistical and econometric approaches have had to be improved and supplemented by machine learning methods. This course discusses a range of analytical methods that may cluster analysis, regression, model selection and regularisation, regression trees, cross-validation, the bootstrap and natural experiments with an emphasis on applications in a data rich environment. The relative strengths and interaction of these methods will be explored across three types of tasks: description, prediction and counterfactual prediction. The distinction between experimental and observational data will be stressed and there may be discussion of less traditional types of data structures such as textual, spatial and network data.