Nutrition data refers to information on the nutritional content of food and its effects on human (and animal) health. This data category also includes information on the effects of processing and storage methods on food.
Doctors, dietitians, and various types of researchers both private and public generate nutrition data. In addition to traditional scientific methods of studying food or diet on human or animal health, researchers increasingly rely on machine learning programs. These ML programs use image recognition systems, spectral imaging systems, and new devices like electronic noses and data techniques like chemometrics and sensometrics.
Since the applications for nutrition data are so wide, the attributes of this data are correspondingly detailed. In other words, the data can reveal information about micronutrients within certain grams of food or about malnutrition level within a nation’s population.
Healthcare industry professionals, public policy makers, NGOs, athletes, veterinarians, ordinary people, and more use this data to improve health. On the individual level, nutrition data enables people to monitor their weight or treat illness. On the population-wide level, the data helps improve the safety of food processing factories or develop more nutritious, easily-grown major crops, and so on.
Certain things make it difficult to determine the quality of nutrition data: first, there is a huge amount of data available, for every application of the subject. Second, machine learning programs using this type of data need long training times before showing use. However, as long as you begin building your dataset with your purpose in mind, the quality should follow. Just ensure you complement the dataset with relevant data from other fields—demographic data for rates of anemia in a region, for example.
Dr. Ibrahim Elali, nephrologist and assistant professor at UConn School of Medicine, created the free smartphone app DecideDiet for people suffering from heart and renal disease. This app provides data about the sodium and phosphorous content in foods in a user-friendly way.
Phosphorous, in particular, is a challenge for DecideDiet as, while the mineral is a common preservative,
the FDA does not require companies to report it on nutritional labels as it does with sodium and potassium.
FAOSTAT Data consists of more than just agriculture and nutrition data; it also consists of commodities trade, land use, food prices, population, and investments data. Users can also search and compare data according to topic (forestry vs aquaculture vs consumer protection, for example), region, or time range.
Gastronomy Databases offers restaurant data for establishments all over the world. With more than fifteen years of experience in gastronomy, their data sets can be filtered by over 100 criteria.