Identification and profiling of socioeconomic and health characteristics associated with consumer food purchasing behaviours using machine learning

Food systems and food-related policies influence food consumption, dietary patterns, and human and environmental health. Consumers play a vital role in enhancing health and sustainability through their purchasing choices. To identify and cluster food purchasing behaviours and map relationships, a cross-sectional survey was conducted across Ireland with a sample size of 957 adults. Two-step cluster analysis, generalised linear models, and recursive partitioning and regression trees were used to elucidate adherence to identified food purchasing behavioural clusters. Three clusters (‘food quality’, ‘taste’, and ‘price’) were identified based on food purchasing priorities and statistically categorised. ‘Food quality’ members were significantly less likely categorically obese (OR = 0.32) and more likely to have a postgraduate degree (OR = 1.59–1.76). ‘Taste’ members were almost twice as likely to be classified as obese (OR = 1.96), have/had diabetes (OR = 2.24), and have secondary-level education as their highest level of attainment (OR = 1.73). ‘Price’ members had the highest mean body mass index (28.03 kg/m²), were more likely younger (25–34 years) (OR = 1.43) and were more likely to have lower annual household income (<€24,999) (OR = 1.89). Machine learning models demonstrated an increasingly efficacious fit for predicting adherence to ‘food quality’ membership (area under the curve = 0.72), with education, body mass index, meat/seafood purchase location, food retailer distance, and dietary pattern identified as major predictors. Findings emphasise the need for tailored, evidence-based policies to modify physical environments, improve economic conditions, and enhance consumer awareness to promote diets balancing nutritional quality and sustainability.