Technical

Stack

The technologies and architecture that power Atlas Amazon, including the frontend, data systems, and geospatial infrastructure.

Atlas Amazon is a web-based spatial platform built from three components: a frontend application, a geospatial layer, and a backend data system. Together they support interactive mapping, structured data, and participatory workflows in a single application.

Frontend

The interface is built with Nuxt (Vue 3 + TypeScript), which keeps maps, metadata, and contribution workflows reactive and in one place rather than spread across separate tools. Nuxt UI supplies the component layer, so the Atlas and this documentation share a single design system.

Geospatial

Spatial data is rendered with MapLibre GL JS, an open-source WebGL mapping library, and vector data is served through PMTiles — a single-file format that delivers large datasets without a dedicated tile server.

That choice has three practical consequences: large spatial datasets stream efficiently, infrastructure stays simple, and the platform scales through static hosting and CDNs rather than server capacity.

Backend

Structured data — metadata, assessments, and user information — is managed in Supabase (PostgreSQL). The backend provides real-time data updates, API-based access, and role-based permissions for contributors, validators, and public users, which is what allows the Atlas to work as a collaborative platform rather than just a data store.

Design Principles

Four principles shape these choices:

  • Open — built on open-source tools and standards
  • Interoperable — compatible with common geospatial formats
  • Scalable — designed to handle large datasets efficiently
  • Modular — components can evolve independently

The aim is long-term sustainability and straightforward integration with external systems.