%20{%20.cls-1%20{%20fill:%20%23fff;%20}%20}%20%3c/style%3e%3c/defs%3e%3cg%20id='Layer_1-1'%20data-name='Layer%201'%3e%3cg%20id='Layer_1-1'%20data-name='Layer%201-1'%3e%3cpath%20class='cls-1'%20d='m220.77,35.74c2.97,0,3.62,1.09,3.41,3.84-1.09,18.77-14.64,37.69-32.18,43.57-5.87,1.96-12.32,2.83-18.48,3.12-12.61.43-25.23.14-37.84.14-1.59,0-3.19.07-4.71.29-6.09.94-10.22,5.65-10.08,11.53.14,5.87,4.78,10.8,10.87,11.09,5.94.29,11.96-.07,17.9.22,13.77.72,23.78,7.54,29,20.59,8.12,20.22-6.6,42.7-27.98,43.06-11.24.14-22.54.07-33.78.07-8.41,0-10.73,1.88-12.69,10.15-.94,3.77-1.74,7.61-3.04,11.31-6.02,17.11-17.9,27.18-35.59,29.07-9.64,1.01-19.5.14-29.79.14,2.25-10,4.35-19.5,6.52-29,1.88-8.19,3.84-16.31,5.58-24.5,3.26-15.01,13.48-23.7,28.49-23.85,13.77-.14,27.47-.07,41.25-.22,5.44-.07,9.64-3.19,11.09-7.97,1.38-4.35-.07-9.79-3.99-12.25-2.1-1.3-4.78-1.96-7.25-2.32-7.83-1.09-15.87-.94-23.41-2.97-19.21-5-31.53-22.54-30.95-42.48.58-19.93,14.35-37.11,33.27-41.46,3.26-.72,6.67-1.09,10-1.09l114.39-.07h0Z'/%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
From Food Labels to Reusable Ingredient Data
How our team built a traceable reference corpus and tested a human-in-the-loop workflow for linking Dutch food ingredients to FoodOn and ChEBI.
Category: Data Science Projects
Articles and insights in the Data Science Projects category.
When Should AI Ask About Your Meal?
For my MSc thesis, I built Snap and Say, a food logging app that uses photos, voice, and selective follow-up questions to balance nutritional detail with ease of use for older adults.
Decoding the Dutch Diet: A Data-Driven Approach to Identifying National Meal Patterns
A deep dive into how I used Association Rule Mining and NLP-powered clustering on national survey data to identify popular Dutch dishes, analyze dietary trends over 15 years, and assess their nutritional quality.
Navigating the Ethical Minefield of DTC Genetic Testing: A Framework for Responsible Innovation
The multi-billion dollar Direct-to-Consumer (DTC) genetic testing industry is an ethical minefield. This post breaks down the key risks—from data privacy to flawed health reports—and proposes a framework for responsible innovation using data justice and co-design principles.
Uncovering Dietary Patterns to Predict Colorectal Cancer Recurrence: A Data-Driven Approach
A case study on using Latent Profile Analysis and Cox Proportional Hazards modeling to identify pre-diagnosis dietary patterns associated with colorectal cancer recurrence. We navigated complex, high-dimensional health data to deliver clinically relevant insights.
Calibrating for Accuracy: Validating a Dietary Questionnaire with Linear Regression
A case study on validating a new Food Frequency Questionnaire (FFQ) for dietary fiber intake. This post details the use of 24-hour dietary recalls as a reference method and the application of linear regression calibration to correct for measurement error, enhancing data accuracy for epidemiological research.