Tokyo Venue Embedding Lab
Building a bias-aware venue dataset for Tokyo bars, restaurants, cafes, and listening rooms before it becomes part of the live map.
Building a bias-aware venue dataset for Tokyo bars, restaurants, cafes, and listening rooms before it becomes part of the live map.
Standardizing how Army watercraft crews at Yokohama North Dock requested maintenance, sourced support, and captured the work.
An exploratory SQL project using maintenance-style records to study delay patterns, recurring failure modes, and readiness signals.
Build log / this website
Define audiences, routes, reusable content types, and page states before writing UI.
Use Codex for layout iteration, copy cleanup, responsive checks, and implementation review.
Build App Router pages, shared cards, desk objects, travel maps, and typed content modules.
Wire Mapbox, Google Places data, Apple Music links, PDF assets, and Vercel project settings.
Run lint, TypeScript, builds, browser checks, map audits, and mobile layout verification.
Commit scoped changes, push the portfolio state, and deploy the verified build through Vercel.
App Router, typed UI, and the shared page system.
Planning, code iteration, copy review, responsive checks, and browser QA.
Map rendering, checked-in place identity and coordinates, and verified Maps links.
Tokyo venue-profile experiments; the public planner uses explicit gates and weights rather than an opaque model.
Optional feedback persistence only when the server-side Firebase configuration is present; otherwise feedback remains local.
Version control, preview-ready deployments, and production release checks.
Links and source material for the media archive and public-safe case studies.