Slimmed Down Nutrient Composition Database

Question

I am working on a project that needs to recommend appropriate foods for people with nutrient deficiencies.

This needs a database with foods and all their nutrient information (All vitamins, magnesium, iodine, zinc, riboflavin etc) I've looked at plenty of massive datasets such as the USDA Food Composition Database and The Canadian Nutrient File.

While these datasets are excellent, they are too large for my work in their current state.

As an example if someone is low on b-vitamins and iron the recommendation would likely be beef. So the data base needs a single option for "beef" (or max 2/3 different cute). The Canada dataset has about 30 with long obscure names such as "ground beef, lean, 0% trimming etc..."

I am looking for a dataset that simplifies all items that are beef and provides it as beef, doing this for all other items.

Manually refining down the 6000 items from the large datasets would take a long time. Does anyone know of a set that meets this slimmed down requirement?

Answer 1

There are two ways to go about this. The first is to average nutrients by food group, and here are US-based datasets already broken out by food groups: http://www.ers.usda.gov/data-products/food-availability-(per-capita)-data-system/.aspx

The challenge with this approach is how to do the averaging - for example, because the 30 types of beef can differ significantly in the nutrients inside them, how will your app decide what exactly to recommend?

Another approach is to start from nutrients, and recommend popular food groups that contribute "a lot of" specific nutrients. This link has datasets about nutrients contributed from each major food group: http://www.cnpp.usda.gov/USFoodSupply-1909-2010