HistoDot — Web Platform for Architectural Heritage Data
Svelte TypeScript Node.js Tailwind CSS PostgreSQL Potree (point-cloud) Docker
Project Overview
Web platform that centralises architectural-heritage datasets, offering search, metadata editing and in-browser 3D point-cloud visualisation. Designed for researchers and conservators to easily explore, annotate, and share large scans.
- Built with Svelte + TypeScript, REST + WebSocket backend in Node.js.
- Integrated Potree so users can load multi-million-point LAS files directly in WebGL, measure distances and add annotations.
- Role-based access control layer for historians vs. public visitors.
Problem & Motivation
Architects and preservation experts often manage gigabytes of scans, photographs, and CAD files, typically scattered across laptops and external drives. I built HistoDot to provide a single, browser-based hub where they can search, filter, and view 3D point clouds without relying on heavy desktop software.
Challenges & Solutions
- Large‑dataset performance → redesigned data storage and indexing to support smooth multi-field search, even at massive scale.
- Real‑time annotation sync → added a WebSocket layer with optimistic locking so teams can annotate simultaneousl without merge conflicts.
- Huge point‑cloud files → pre‑processed uploads into small view‑dependent tiles and loaded them on demand, keeping navigation fluid in the browser.
Key Takeaways
- Tiling and on‑demand loading keep massive 3D scans interactive on ordinary laptops.
- Web-based access removes the need for specialized desktop tools, making the project more inclusive and easier to adopt.
- Role‑based workflows let institutions protect master data while still crowd‑sourcing annotations.