Project

Data analytics for food chains and consumer-oriented research

Agrifood research is rapidly adopting digitalized, data-driven methods. This requires researchers to share data, models and tools. This project aims to establish and evaluate an effective research infrastructure for consumer and food safety research.

In this project we work with four use cases in agri-food.

  • Creating evidence-based and personalised dietary guidance.
  • Automatic collection of data on food fraud and food safety.
  • Surveillance of farms and slaughter houses on zoonotic pathogens.
  • Developing harmonized and efficient consumer surveys.

 

The design of the Metadata Library for models and datasets has been established. A core functionality of any research infrastructure is to support researchers in finding datasets and models created by others. In line with the FAIRification guidelines, the selected metadata fields are mapped to standard vocabularies. 

The use case on digital dietary advice aims to develop a demonstrator. Given input on consumer profile, life style, allergies, the system generates an healthy and acceptable alternative diet. We are discussing with two large companies how to implement the consumer advice research tool in practical applications.

We have improved the food fraud model and added three new data sources, based on literature search. Automatic data retrieval using open source software KNIME is realised for three European databases. The industry is interested in automated data retrieval and the updated food fraud model.

We have started developing machine learning models to design data-driven selection of pig farms for surveillance  in slaughter houses. We focus on the monitoring programme for antibiotic resistance (samples taken at slaughter houses) where we share data with NWVA and RIVM.

A consumer survey tool has been set up for standardized questions and answers to assess consumer related issues. The tool supports semi-automatic generation of a questionnaire (via smartphone or desktop solution). We have explored the link between the TKI PPS Smart Food Intake and the service of the easy questionnaire, e.g. by adding standard questions to SFI tools (Unilever, Philips, Danone and Friesland Campina).

We have presented the results to industrial companies an international conferences.

Publicaties