Categories taxonomy alternatives

From Open Food Facts wiki
Revision as of 10:29, 9 November 2023 by Aleene (talk | contribs) (10000219)
Jump to navigation Jump to search

The OFF categories taxonomy for products is certainly not the only around. There are many alternatives. But none of these alternatives fit the OFF needs. The ones we know of will be reviewed below. Any of the alternatives should fit the following requirements:

  • Comprensible for the end-user: anyone should be able to set the category for a product. It is similar to what a consumer encounters the the aisles of a shop;
  • Agribalyse match: in order to be able to calculate the environmental impact a mapping to the Agribalyse categories must be possible;
  • AOP match: some products have an AOP-label which corresponds to a single category. Any taxonomy must be detailed enough to accomodate this;
  • Diet hierarchy: the hierarchy of a taxonomy should be able to express many diets (vegan, etc), so that inheritance can be applied;

Agribalyse

Ciqual

GPC

GPC contains a ery large categorisation of products (food and non-food). The categorisation contains a hierarchy (based in what criteria?). The granularity of the hierarchy is reasobly specific, but falls short of the OFF taxonomy. I had look at the GPC categories (https://gpc-browser.gs1.org), whether mapping to our categories would be feasible. An issue that I already see is that our tree can be more specific. Sometimes we have details not expressible in GPC. GPC uses multiple attributes instead of children. For instance the grade of an oil is an attribute, whether it is organic is an attribute and the type of oil is an attribute. This implies that some of our categories need multiple attributes AND corresponding values. GPC is already used by some producers who upload their data. We could use it to assign products to categories, and as additional check. As no attributes are passed by producer we can skip these for the moment.

Differences

There are some differences when trying to map GPC to the Categories taxonomy:

  • Broader: the GPC is sometimes broader, examples:
    • Nuts/Seeds - Prepared/Processed (Out of Shell) (10000236)
  • Process-specific: The GPC emphasises a specific process for a category. Examples:
    • Shelled/Unshelled in Nuts/Seeds - Prepared/Processed (Out of Shell) (10000236)
    • Perishable/Shelf stable in Fruit Juice - Ready to Drink (Perishable) (10000219)

WHO