A Database To Make Sense Of Ultra-Processed Food

‘Food processing’ refers to anything we do to transform our food after it has been harvested or killed. Most of the food that we buy is processed in some way, but some foods are much more processed than others. Food components are transformed, enriched, extracted or added and mixed together to create compositions that would never be found in nature. These are known as ultra-processed foods.

Such foods appear to be harmful to our health for a variety of potential reasons that you can read about in this article. One of the more surprising findings of recent research has been that foods that have undergone more processing appear to be worse for you than less processed foods containing exactly the same ingredients. For that reason, regardless of what you are eating, it may be worth seeking out the less processed version of that food.

But how are you supposed to tell which of two frozen pizzas is more processed? Simply looking at the list of ingredients isn’t going to help much, as you still need to know how processed those ingredients are, which means you need to know how processed those ingredients’ ingredients are…

That’s where this interesting project called GroceryDB comes in. This is an online database that estimates the level of food processing using machine learning. Though it only contains product sold in the United States, it still makes some interesting browsing and is freely accessible at truefood.tech. 

How does it work?

GroceryDB is powered by FPro, a food processing score calculated by a machine learning algorithm based on a product’s ingredients. When browsing, you will see a number between 1 and 100 for each product, which refers to its percentile position on the FPro range for that category. For example, if a pizza has a score of 5, then 95 out of 100 pizzas will be more processed than that, making this a relatively unprocessed product (for a frozen pizza).

The database also provides nutritional content and an ingredient tree. This shows a breakdown of all of the ingredients and additives that went into the product.