Calculating and Presenting CO2e Emissions of Procured Food: Automating classification of procured food and presenting calculations in an organisational context

Calculating and Presenting CO2e Emissions of Procured Food

Nielsen, Søren Støvring ; Dam, Bjørn Daniel ; Hein, Magnus Helsvad

4. Term

Digitalisation and Application Development, MSc

2022

2022-06-10

249

Dette projekt undersøger, hvordan Madservice Aalborg kan bruge et digitalt værktøj til at spore klimabelastningen (CO2e) fra indkøbte fødevarer. Sporing forventes at blive obligatorisk, men det er stadig uklart, hvordan opgaven skal løses. Vi gennemførte interviews, en workshop, prototyping, en evaluering og en tematisk analyse for at forstå organisationen og definere kravene til løsningen. Vi anvendte også maskinlæring til automatisk at matche fødevarer fra fakturaer med poster i The Big Climate Database, som giver klimaoplysninger om fødevarer. Maskinlæring kan gøre beregningerne mere ensartede og gennemførlige, men nøjagtigheden afhænger af flere faktorer. For bedst at støtte mere bæredygtige indkøb bør løsningen præsentere information på måder, der passer til brugerne, deres kontekst og formålet med dataene. Projektet viser eksempler på visuelle design, der indarbejder vigtige kontekstuelle hensyn i grænsefladen.

This project examines how Madservice Aalborg can use a digital tool to track the carbon footprint (CO2e) of the food it buys. Tracking is likely to become mandatory, but how to do it is still unclear. We carried out interviews, a workshop, prototyping, an evaluation, and a thematic analysis to understand the organization and define the requirements for the tool. We also applied machine learning to automatically match food items listed on invoices to entries in The Big Climate Database, which provides climate information about foods. Using machine learning can make calculations more consistent and practical, but accuracy depends on several factors. To effectively support more sustainable purchasing, the tool should present information in ways that fit users, their context, and the purpose of the data. The project offers example visual designs that show how important contextual considerations can be built into the interface.

[This summary has been rewritten with the help of AI based on the project's original abstract]

Food sustainability ; HCI ; Machine Learning ; Prototyping ; Information Visualization

Download
View record in AAU Student Projects