This research aims to combine computer science techniques, such as machine learning, with psychological literature in order to improve the prediction of behaviour in big data sets. I am working in Partnership with Boots and Walgreens Alliance. We have linked 15,000 personality questionnaire responses to 10-years’ worth of Boots loyalty card transaction data. My PhD research looks to answer questions in 3 areas:

1. Methods: Can psychological scientific investigation benefit from using machine learning? For example, a) can non-linear machine learning methods help tell us something about the data that traditional statistics used in psychology cannot? And b) can prediction accuracy as a metric, and cross-validation as a method, help to overcome some of the over-fitting problems which occur when using variance explained on a single data set as the main measure of the model’s success?

2. Theory: How can data inform the psychological literature, and even generate new theories? It is rare in Psychology to have such a large combined psychometric and behavioural dataset. This research will explore a bottom-up approach to the psychological mechanisms that predict behaviour. Using powerful machine learning algorithms on millions of data points it is possible to identify novel predictors of behaviour which have not previously been recognised or explained by psychological theory.

3. Applications: In the competitive business and marketing world, how can personality traits improve prediction of consumer behaviour? Boots and Walgreens Alliance are looking to gain a greater understanding of their customer base in order to carry out better target marketing, gain insights into their sales reports, and to inform the development of new products. Using individual difference measures and psychological theory on personality traits, life events, and the mechanisms predicting behaviour will help Boots to understand their customer’s behaviour, desires, and needs – thus improving the prediction of future customer behaviour.