Aussie Adventure — Travel AI meets Fashion

Project Overview

Aussie Adventure is a travel–fashion prototype that turns landscape photos into wearable palettes. It extracts 5 key colors and high-level scene tags from destination images, then blends activity (e.g., walking, hiking, dining) with weather to propose 2–3 concise outfit ideas. The goal is speed and delight: get a “looks-right” suggestion in seconds, with one-line explanations that feel human, not robotic. The UI favors quick tweaks—swap a color, change an activity, or toggle layers—so users can nudge results without starting over. The system is intentionally lightweight, designed to be embedded into a trip planner or a retailer’s inspiration page.

What I Did

• Drove scoping and user research; mapped how travelers currently plan outfits for multi-activity days and weather uncertainty.
• Designed the end-to-end UX: upload → palette extraction → activity & weather constraints → 2–3 outfits → one-line explanations.
• Implemented a simple rules core that mixes palette mood with practicality (breathability, layers, shoes), avoiding heavy taxonomies.
• Built a clean Next.js + Tailwind interface with chip-based controls and a lightweight editor for color swap / activity switch.
• Authored explanation templates with tone control; kept copy under ~140 chars so it reads like a friendly stylist, not a spec sheet.
• Set up quick user-style pilots to test timing, clarity, and satisfaction; iterated on defaults for color prominence and layer choice.
• Outlined integration paths for ‘shop the look’ (e-commerce APIs) and feedback loops (accept/reject) to personalize over time.

Reflection

Turning photos into palettes is delightful, but guardrails matter. Simple, readable tags beat deep hierarchies when the goal is fast inspiration. One crisp line of “why this works” consistently outperformed long paragraphs in user feedback, and small editor affordances (swap a color, toggle layers) made the system feel collaborative rather than prescriptive. If I extend this, I would: (1) plug in e-commerce APIs for instant “shop the look”, (2) learn from accept/reject signals to tune palette prominence and tone, and (3) add a packing-list view that rolls daily outfits into a compact, weather-aware set of items.