Identifying Belgian protein consumption typologies by means of clustering and classification to move towards personalized advices for sustainable and nutritious food choices

Abstract

Protein sources are determinant for the environmental impact and nutritiousness of our food consumption patterns. Dietary advices, in terms of nutritional values and, more recently, the associated environmental impacts, are mostly formulated at nationwide level. However, actual food consumption patterns are variable between individuals within nations, leading to a need for more personalized dietary advices to ensure the feasibility of potential dietary changes. This research addresses the variability within Belgian food consumption patterns by identifying different “protein consumption typologies”: consumer groups with different consumption frequencies of different protein sources. Hereto, both statistical clustering and classification were applied to food consumption data, the latter using predefined criteria related to meat consumption. The obtained typologies were further analyzed based on personal characteristics and compliance to nutritional guidelines. Five clusters and five classes were identified, with each survey participant belonging to one cluster and to one class. The clusters differed mainly in milk- and grain product intake, while the classes differed in the intake of meat products from the onset. Both clustering and classification showed that only a small part of the Belgian population frequently adopts plant-based protein sources, that meat products are predominant and that compliances to dietary guidelines are generally low. The typologies are the starting point for more personalized dietary advices to lower environmental impacts while ensuring adequate nutritional value, moving away from nationwide advices.