nate tran
Independent mind.
Multidisciplinary practice.
← Selected work / Explorite

Research became
a real product.

A travel-discovery concept evolved into a live PWA with bookings, reviews, messaging, and a research-backed crew voting flow.

UX research / Product design / Full-stack PWA
03 / UX research + interface design

Less planning.
More living.

Original Explorite mockup showing the designed discovery interface on an iPhoneexplorite ↗
01

The original problem

You land somewhere new with two free days, and every option is either a tourist trap with 40,000 reviews or a listicle written for someone else. Travelers need to find good local activities quickly and trust what they find. On a short trip, a bad pick is expensive.

02

Research first

I ran and recorded user interviews, then took the raw input through affinity diagramming, empathy mapping, and journey mapping. Two research-based personas kept later decisions accountable to real traveler types instead of my own preferences.

The first prototype covered the core loop across six screens: discover, evaluate, and coordinate. It proved the research-to-interface process before the product was functional.

03

The research found the real differentiator

The journey map showed the experience only dipped negative at one stage: checking availability with other people. Discovery was not the only problem. Social coordination was where the trip lost momentum.

That finding became the product’s distinguishing flow. A trip owner proposes two to four times, then friends vote In, Maybe, or Out through a share link with no account required. The feature exists because the research pointed to it.

04

From prototype to live PWA

In August 2026, I rebuilt Explorite with a real Supabase backend and a new Field Journal design system: warm paper, editorial type, restrained terracotta, and documentary photography.

Real accounts and reservations

Email and password authentication, confirmed bookings, ticket codes, and cancellation.

Trust that can be checked

Seeded reviews, with a server-side rule that only marks a review verified when the writer has a confirmed booking.

Crew coordination

Share-token voting works without an account while invalid tokens return no plan data.

Host messaging

Messages use live delivery, with contextual host replies generated behind the scenes.

Strictly separated data

The app shares infrastructure with this portfolio’s backend, but namespaced tables, explicit grants, and row-level security keep their data isolated.

Use the live product
05

Outcome

Explorite now shows both sides of the work: the original research process and the judgment required to turn one important finding into a real, secure, usable product. The prototype did not disappear. It became the evidence behind what shipped.

06

What I’d do next

  • Usability-test the coordination flow with real travel groups.
  • Measure whether verified reviews and attribution improve trust at decision time.
  • Refine the experience catalog based on real bookings rather than seed data.
Next projectStella Booth