nyc-mesh

Python toolkit for converting NYC open 3D source data into web-ready geodata.

Authored by Blaise Albis-Burdige.

nyc-mesh focuses on the messy middle between raw public CityGML releases and practical outputs for browsers, notebooks, and reproducible analysis workflows.

What ships in the current line

The current package now includes a real official-data workflow:

• load local or zip-wrapped official CityGML, LiDAR, DEM, and footprint inputs
• fetch official PLUTO and building-footprint context for a chosen study-area bbox
• reproject CityGML source coordinates from EPSG:2263 to EPSG:4326
• clip buildings to named study areas or explicit WGS84 bounding boxes
• join PLUTO-style attributes onto extracted buildings
• generate lightweight terrain meshes from official DEM or LiDAR inputs
• export GeoJSON, GeoParquet, glTF, and a minimal 3D Tiles package
• build typed cache manifests for real study areas

Why this exists

NYC publishes unusually rich 3D geospatial data, but the raw formats are hard to use in practice. The source files are large, specialist, and awkward to transform into lightweight outputs for web rendering or neighborhood-scale analysis.

This package aims to make the first useful workflow feel like:

1. point at a CityGML source file
2. extract and clip the buildings you care about
3. export web-friendly artifacts or analysis-ready files

Quickstart

Install:

pip install nyc-mesh

Export GeoJSON from a real CityGML source:

nyc-mesh export-geojson --input "C:/path/to/DA_WISE_GML.zip" --output buildings.geojson

Examples

examples/ now follows the same self-contained project pattern used by nyc311. Start with:

• examples/quickstart-citygml/
• examples/landmark-3d-stack/
• examples/building-height-analysis/
• examples/example-template/

Python example

from pathlib import Path

from nyc_mesh import models, pipeline

pipeline.export_citygml_geojson(
    Path("C:/path/to/DA_WISE_GML.zip"),
    Path("buildings.geojson"),
    bbox=models.BoundingBox(
        min_lat=40.70,
        min_lon=-74.02,
        max_lat=40.72,
        max_lon=-73.99,
    ),
)

Current assumptions

The official-data workflow is intentionally opinionated:

• large CityGML / DEM / LiDAR archives stay out of git and are treated as local cache assets
• only buildings with bldg:measuredHeight
• source coordinates are treated as EPSG:2263
• outputs are reprojected to EPSG:4326
• optional clipping uses a WGS84 bounding box

PLUTO joins, real footprints, and terrain inputs are treated as practical building blocks, while the examples document exactly which official sources or local cache paths they need.

Documentation

• Hosted docs: nyc-mesh.readthedocs.io
• Local preview: make docs
• Strict docs build: make docs-build

Quick links

Docs: Home, Getting Started, CLI Reference, Pipeline Guide, Architecture, Examples, Python API, Contributing, Releasing, Changelog

Development

make install-dev
make test
make lint
make docs-build
make ci

License

MIT.