Hi everyone,
I’ve been working on a personal project called SHINOBI EARTH, a fictional 3D geospatial environment that I’m using as a sandbox for experimenting with CesiumJS, WebGIS, spatial analysis, and digital-twin concepts.
The application is built with:
- React
- TypeScript
- CesiumJS
- Vite
- Custom geospatial datasets
- Procedural 3D geometry
- Client-side spatial analysis
The current prototype includes:
- 1,200 spatial analysis cells
- 882 buildings
- Custom terrain and 3D scene geometry
- Roads, rivers, vegetation and landmarks
- Building-density analysis
- River-proximity analysis
- Flood-risk heuristic overlay
- Mountain/forest classification
- Expansion-suitability analysis
- Data/renderer reconciliation and validation
One of the things I focused on was keeping the analytical dataset and the Cesium visualization consistent. I implemented validation checks to make sure the generated building data, spatial grid assignments, and rendered entities remain synchronized.
The project also includes performance work around geometry generation and analytical overlay construction.
The environment is fictional and inspired by a ninja-world setting, but the purpose is primarily technical: using a visually interesting environment as a sandbox for exploring 3D GIS and digital-twin workflows.
Live Demo:
Source Code:
I’d really appreciate feedback from other CesiumJS developers.
In particular, I’m interested in feedback on:
- Cesium scene architecture
- Entity vs. primitive usage
- Large-scale 3D geometry performance
- Spatial-analysis visualization
- Terrain and custom geometry approaches
- Further ways to optimize the rendering pipeline
Thanks for taking a look!