Nutritional epidemiologist have used for long time traditional food-intake questionnaires, which are usually time-consuming, therefore imposing a tight trade-off between the depth of the survey and the size of the cohort. In the last years, thanks to recent technological advances such as the large diffusion of mobile phones, food-related studies have been involving more and more digital approaches. In order to obtain a fast recording of food intakes, tools like mobile apps would definitely benefit from open databases, that would allow to match a food product's barcode or picture to its nutritional content. Here, we first present the workflow behind FoodRepo [1], an open repository of barcoded food products, including the steps for its data validation, long-term maintenance and international expansion. FoodRepo currently contains more then 24,000 barcoded products, collected in three of the main Swiss food retailers. For each item, FoodRepo contains the information reported on the food package, such as the product name, ingredients and nutrition facts, all data that can be programmatically accessed through our open API (Application Programming Interface). Second, we present the analysis that allowed to assess the nutritional diversity of the Swiss food market, thanks indeed to the data contained in FoodRepo [2]. In particular we used predefined algorithms that measure the nutritional values of food products, in order to compare different food types and common dietary restrictions. We furthermore applied clustering and network analysis to identify relevant products and ingredients communities. These first results, together with the planned international expansion of FoodRepo will likely lead to a deeper and broader understanding of the nutritional diversity of the food available in supermarkets, hopefully a first step towards a new type of nutrition, digital and personalized.
