Thesis subject

Protein Quality of online Vegan Recipes (MSc)

The transition towards plant-based diets is increasingly recognized for its potential benefits on health and the environment. However, the nutritional quality of vegan recipes available online and their popularity among consumers remains an area requiring in-depth analysis. This research aims to employ advanced data science methods to scrape a vast array of vegan recipes from the internet, assess their Meal Protein Quality Score (MPQS), and analyze various factors predicting the protein quality.

Short description

The transition towards plant-based diets is increasingly recognized for its potential benefits on health and the environment. However, the nutritional quality of vegan recipes available online and their popularity among consumers remains an area requiring in-depth analysis. This research aims to employ advanced data science methods to scrape a vast array of vegan recipes from the internet, assess their Meal Protein Quality Score (MPQS), and analyze various factors predicting the protein quality. By integrating nutritional analysis with popularity indices such as user ratings, social media shares, and comments, the study seeks to identify key characteristics of high-quality, well-received vegan recipes. This approach not only contributes to nutritional science and dietary recommendations but also provides insights into public dietary preferences and behavior towards plant-based eating.


Objectives

  1. To develop a comprehensive dataset of vegan recipes by scraping data from multiple online sources.
  2. To calculate the MPQS of these recipes based on a set of nutritional parameters aligned with dietary guidelines.
  3. To identify vegan meals with good protein quality scores
  4. To employ machine learning techniques to identify predictors of high nutritional quality among vegan recipes, such as combinations of specific ingredients.
  5. To provide evidence-based recommendations for nutritionists, dietitians, and the public on identifying high-quality vegan recipes.

Tasks

The work in this master thesis entails:

  • Data Collection: Utilize web scraping techniques to collect a large dataset of vegan recipes, including ingredients, nutritional information, and popularity metrics (e.g., likes, shares, comments) from various websites and social media platforms.
  • Nutritional Analysis: Apply the MPQS framework to assess the protein quality of each recipe.
  • Identification Analysis: Use data science methods to analyse the data, identify patterns, and determine significant predictors of recipe quality.
  • Evaluation and Validation: Validate the models using a separate test dataset and assess their accuracy and reliability in predicting the nutritional quality and popular of vegan recipes.

Literature

  • Pol Grootswagers, Sine Højlund Christensen, Marielle Timmer, William Riley, Lisette de Groot, Inge Tetens, Meal Protein Quality Score: A Novel Tool to Evaluate Protein Quantity and Quality of Meals, Current Developments in Nutrition,Volume 8, Issue 9,2024,104439, https://doi.org/10.1016/j.cdnut.2024.104439.
  • · Hertzler SR, Lieblein-Boff JC, Weiler M, Allgeier C. Plant Proteins: Assessing Their Nutritional Quality and Effects on Health and Physical Function. Nutrients. 2020 Nov 30;12(12):3704. https://doi.org/10.3390/nu12123704.

Requirements

  • Courses: FOR INSTANCE: Programming in Python (INF-22306), You can explain the main biochemical processes involved in digestion, absorption and cellular metabolism of macro-nutrients. (based on Nutritional Physiology)
  • Required skills/knowledge: data scraping, statistical skills (machine learning), report writing

    Key words: Protein Quality, Machine learning, Data scraping, vegan recipes.

    Contact person(s)

    Yamine Bouzembrak (yamine.bouzembrak@wur.nl)
    Pol Grootswagers (pol.grootswagers@wur.nl)