Thesis Defence: HealKitchen—A Persuasive, AI-Driven Mobile Application to Influence Healthy Dietary Habits

Iyanuoluwa Sowande, supervised by Dr. Ifeoma Adaji, will defend their thesis titled “HealKitchen—A Persuasive, AI-Driven Mobile Application to Influence Healthy Dietary Habits” in partial fulfillment of the requirements for the degree of Master of Science in Computer Science.

An abstract for Iyanuoluwa Sowande’s thesis is included below.

Defences are open to all members of the campus community as well as the general public. This defence will be offered in hybrid format.  Registration is not required to attend in person; however, please email ifeoma.adaji@ubc.ca to receive the Zoom link for this defence.

---

Abstract

Unhealthy dietary habits remain a global concern, particularly among individuals managing chronic health conditions. While many mobile nutrition applications exist, they often emphasize calorie counting or weight loss, overlook personalization, and lack a foundation in behavioural change theories. This research presents HealKitchen, a persuasive mobile application designed to support healthier dietary behaviours by leveraging behavioural science, artificial intelligence, and user-centred design.

This research addresses five main objectives: the development of a mobile health app to influence healthier dietary choices, the implementation of behavioral theories into the mobile application, the implementation of a recommendation system that delivers personalized nutritional guidance, the development and evaluation of an object detection model capable of identifying grocery items, and the assessment of the mobile application’s usability, perceived usefulness, and engagement.

HealKitchen integrates a range of personalized features, including a recipe recommendation system adapted to user health requirements, a recipe curation engine, and an image recognition model that identifies grocery items and provides nutritional information.

The application design draws on five foundational behaviour change theories, operationalized into persuasive features that nudge users toward healthier choices.

To evaluate the application, a mixed-methods user study involving 20 participants was conducted. Quantitative survey data revealed a high level of usability (mean = 4.6), perceived usefulness (mean = 4.45), and aesthetics (mean = 4.49). Thematic analysis of qualitative feedback highlighted the need for recipe diversity, improved multimedia support, and additional features, informing iterative design enhancements. A second study, pending ethics approval, will evaluate the app’s long-term behavioural impact using A/B testing.

This research demonstrates the feasibility of building a persuasive and theory-driven mobile health application for personalized nutrition support.