Applying Natural Language Processing to U.S. Food Retail Data to Categorize Foods by Level of Processing

Mariah Ehmke, Abigail Okrent, Koroles Awad, Jill McCluskey and Brandon Restrepo

No 404883, 2026 Annual Meeting, July 26 - 28, 2026, Kansas City, Missouri from Agricultural and Applied Economics Association

Abstract: Differentiating food products by level of industrial food processing is essential to measuring relationships among food processing, food demand, and public health. The Circana (formerly known as IRI) scanner data lacks identified measures of food product processing, but contains other food product descriptors that can be used to predict the processing of food products. This report specifies procedures to use Natural Language Processing with a Naïve Bayes model in Python’s scikit-learn to efficiently classify retail food purchases by level of food processing according to the NOVA food classification system. This method accurately predicts NOVA classification 94 percent using only the food item descriptions as predictors. Based on this classification we find 60 percent of retail food sales to be ultraprocessed in 2023, an increase from 55 percent in 2020.

Keywords: Food; Consumption/Nutrition/Food; Safety (search for similar items in EconPapers)
Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://ageconsearch.umn.edu/record/404883/files/AAEA_July82026.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:ags:aaea26:404883

DOI: 10.22004/ag.econ.404883

Access Statistics for this paper

More papers in 2026 Annual Meeting, July 26 - 28, 2026, Kansas City, Missouri from Agricultural and Applied Economics Association Contact information at EDIRC.
Bibliographic data for series maintained by AgEcon Search ().