<ns4:p>There is a lack of datasets in Colombia that characterize the nutritional components and other similar information about food items. This study describes a structured process to develop datasets that captures the preferences and purchases of food items by a selected group of people. The datasets would classify products according to their sodium and sugar content. The outcome of this structured process would include three datasets, each with a different focus: the first contains data on food preferences, the second contains the purchase history according to the invoices obtained, and the third contains characteristics of the food items such as its brand, category, sodium and sugar content levels, among others.</ns4:p>