Project Hodo
AI-powered trip recommendations built around who you are as a traveler, not where everyone else went.
Focus: AI & Travel Psychology
Video Trailer
Project Hodo — Thesis Trailer
Video coming soon — add a YouTube or Vimeo embed URL tovideo_url in the thesis markdown file.
Abstract
Hodo is a mobile application that reimagines travel discovery through the lens of travel psychology. Rather than surfacing popularity rankings, Hodo begins with a short psychometric assessment that identifies the user's travel personality across six dimensions—adventure-seeking, cultural immersion, social connection, restorative escape, sensory experience, and creative exploration. A fine-tuned language model then generates personalized destination recommendations, day-by-day itinerary structures, and experience suggestions calibrated to those personality dimensions.
Design Challenge
Mainstream travel apps optimize for aggregated popularity. The result is a homogenization of travel—everyone visiting the same landmarks, the same cafés, the same Instagram backdrops. Hodo argues that better travel starts with better self-knowledge, and that AI is well-suited to bridge the gap between who a traveler is and what would genuinely delight them.
Technical Approach
Personality assessment uses a validated 24-item instrument derived from travel motivation research. User responses are encoded as a personality vector and passed to a GPT-based model fine-tuned on curated travel writing. The app was built with React Native; the recommendation backend runs on a Python FastAPI service deployed to Render.
Outcomes
User testing with 15 participants showed that 87% found Hodo's recommendations "more aligned with what I actually want" compared to a control condition using a leading travel app. The project was presented at the IxDIA Spring Exhibition 2022.
Process Documentation
[Add personality framework diagrams, UX research, wireframes, and final prototype screens here]