geoenv¶

Map geometries to environmental semantics

Release v0.5.0. (Installation)

geoenv is a Python library that maps geospatial geometries, such as points and polygons, to environmental terms in vocabularies/ontologies (e.g. ENVO). It’s like reverse geocoding, but for environments.

Features¶

Broad scale environmental context: Provides consistent broad scale environmental context supplementing local scale environmental descriptions.
Global Coverage: Provides worldwide resolution of terrestrial, coastal, and marine environments.
GeoJSON Output: Outputs data as a GeoJSON Feature, for integration with other tools and libraries.
Concurrent Data Resolution: Leverages asyncio to query multiple geospatial data sources concurrently, providing fast results.
Modular and Extensible: Designed with a modular architecture to facilitate integration of new data sources and vocabularies.

Quick Start¶

Install from PyPI:

$ pip install geoenv

Resolve a point location to environmental descriptions:

import asyncio
from geoenv.geometry import Geometry
from geoenv.resolver import Resolver
from geoenv.data_sources import (WorldTerrestrialEcosystems,
                                 EcologicalMarineUnits,
                                 EcologicalCoastalUnits)

# Define a geometry in GeoJSON format (Point or Polygon)
geometry = Geometry(
    {
        "type": "Point",
        "coordinates": [
            -122.622364,
            37.905931
        ]
    }
)

# Set up the resolver. When the location's environment is not known,
# multiple data sources are included to cover potential environment
# types.
resolver = Resolver(
    data_source=[
        WorldTerrestrialEcosystems(),
        EcologicalMarineUnits(),
        EcologicalCoastalUnits(),
    ]
)

# Resolve the geometry to environmental descriptions. The resolver
# queries multiple data sources concurrently using `asyncio`.
response = asyncio.run(resolver.resolve(geometry))

# Access response data.
print(response.data)

The response is a GeoJSON Feature with environmental terms mapped to ENVO (by default). Only resolved environments are included:

{
  "type": "Feature",
  "identifier": null,
  "geometry": {
    "type": "Point",
    "coordinates": [
      -122.622364,
      37.905931
    ]
  },
  "properties": {
    "description": null,
    "environment": [
      {
        "type": "Environment",
        "dataSource": {
          "identifier": "https://doi.org/10.5066/P9DO61LP",
          "name": "WorldTerrestrialEcosystems"
        },
        "dateCreated": "2025-03-07 15:53:09",
        "properties": {
          "temperature": "Warm Temperate",
          "moisture": "Moist",
          "landCover": "Cropland",
          "landForm": "Mountains",
          "climate": "Warm Temperate Moist",
          "ecosystem": "Warm Temperate Moist Cropland on Mountains"
        },
        "mappedProperties": [
          {
            "label": "temperate",
            "uri": "http://purl.obolibrary.org/obo/ENVO_01000206"
          },
          {
            "label": "humid air",
            "uri": "http://purl.obolibrary.org/obo/ENVO_01000828"
          },
          {
            "label": "area of cropland",
            "uri": "http://purl.obolibrary.org/obo/ENVO_01000892"
          },
          {
            "label": "mountain range",
            "uri": "http://purl.obolibrary.org/obo/ENVO_00000080"
          }
        ]
      }
    ]
  }
}

Motivation¶

Finding datasets based on their environmental context is a challenge in data synthesis. The process often relies on vague or inconsistent metadata. This variability presents a barrier to reliable, large-scale analysis due to time lost in data discovery and incomplete search results.

geoenv helps address this challenge by using a dataset’s originating location as a consistent and objective starting point. It can programmatically map the geometry of this location to standardized environmental terms, providing a scalable and repeatable method for generating interoperable metadata. This approach aims to enrich datasets with uniform, semantic metadata, making them potentially easier to discover, query, and integrate at scale.

Contributing¶

We welcome contributions! If you know of a useful data source or vocabulary, and have ideas for new features, or find a bug, please open an issue to start a discussion.

License¶

This project is licensed under the terms of the MIT license.