Comparing the environmental impacts of recipes from four different recipe databases using Natural Language Processing

The calculation of environmental impacts from recipes remains a barrier to effective uptake of sustainable diets. In our project, we use pilot digital humanities methods to explore digitised recipe texts from websites in English, Dutch and German. Using the natural language processing toolkit GATE [1], we have developed customised tools to automatically extract ingredients, quantities and units from 220,168 Indexed recipes and match them to a food environmental impact database of 4500 ingredients (using the classification system FoodEx2). This database, based on environmental data from Poore and Nemecek (2018), provided Land Use (m2/FU), GHG Emissions (kg CO2eq/FU, IPCC 2013 incl. CC feedbacks), Eutrophying Emissions (g PO43-eq/FU, CML2 Baseline), Stress-Weighted Water Use (L/FU), and Freshwater Withdrawals (L/FU) for each ingredient. This allowed the calculation of these impacts at the mean, 5% and 95% confidence level per recipe and per portion. This has enabled us to explore the environmental impacts of vegan, vegetarian and non-vegetarian recipes if we were to cook these recipes using contemporary ingredients. To validate this tool we manually calculated the impacts of 50 recipes from 4 websites BBC Good Food, Albert Heijn/Allerhande, AllRecipes.com (Trattner et al 2017) and Kochbar (Trattner et al 2019) and compared these to the results from our tool