GeoHearth combines data from multiple independent sources. These sources differ in geographic coverage, spatial resolution, observation date, update frequency and methodology. A high-resolution map interface does not increase the underlying resolution of a dataset.
Current sources
| Provider / source | Use in GeoHearth | Important limitation / attribution |
|---|---|---|
| OpenStreetMap contributors | Mapped roads, places, buildings, water, and other geographic features. | © OpenStreetMap contributors. Coverage and completeness vary by area. |
| Nominatim / Photon or configured geocoder | Place search and reverse geocoding where enabled. | Subject to provider usage policies and rate limits. Search results may be incomplete. |
| CARTO | Basemap tiles where configured. | Attribution must remain visible where required. |
| Esri World Imagery / water data | Satellite or mapped water information where configured. | Imagery date, mapped coverage and resolution vary. Required attribution must remain visible. |
| Wikimedia Commons | Nearby geotagged context photos, where available. | Each displayed photo links its author, licence and original source, with the reported photo date and distance from the selection. Photos may be historical and do not establish current environmental conditions. Proposed-direction drawings are concept illustrations, not forecasts or approved site designs. |
| NASA GIBS / MODIS | Satellite imagery and land-surface-temperature context where configured. | Clouds, observation timing and coarse pixel sizes affect interpretation. |
| NASA Earth Observatory | Environmental context/news links. | Contextual information; not proof of a local condition. |
| Open-Meteo / Copernicus CAMS | Weather and modeled air-quality information. | Modeled estimates are not street-level certified monitors. |
| ARPA Lombardia / Regione Lombardia | Municipality-wide daily PM2.5 and PM10 estimates for matching Lombardy localities, from the official sensor catalogue and municipal-estimate dataset. | Monitoring-informed model estimates, provisional and retrospective; not live street measurements or a US AQI. The date is shown. Stale or unavailable values are withheld and the regional model remains explicitly labelled as fallback. |
| USFS / NLCD Tree Canopy Cover | Tree-canopy coverage where available. | Coverage may be regional and dataset years may differ from current conditions. |
| Meta / World Resources Institute Canopy Height Maps v2 | Global modeled canopy-height GeoTIFF tiles from the public AWS Open Data archive. Used as a global fallback for the canopy map and sampled-area tree-height estimate, including India. | The raster is approximately 1.19 m at source, but the app resamples map views and area samples for performance. For larger selections, only a central sample up to 2 km on each side is analyzed; its share of the selected area is reported. Colors represent modeled tree height, not canopy crown-area density. Underlying imagery dates vary by place and pixel; local model error and missing tiles remain possible. Licensed CC BY 4.0. AWS data and license · WRI dataset notes. |
| Altimeter Cloud Tree Canopy API · Meta/WRI CHMv2 | High-resolution mapped tree-height cells and an estimate of the share of covered cells with trees at least 3 m tall; used for the selectable canopy-height overlay and as a regional canopy-data source. | Source imagery is mainly from 2017–2020. The API covers only mapped regions; coverage and sampled-area limits are shown. This tree-height proxy is not canopy crown-area percentage, a current street-tree inventory, or proof that transparent cells have no vegetation. Data credit: AltimeterCloud.com; underlying Meta/WRI and ETH data licenses also apply. API and attribution details · Meta/WRI dataset. |
| OpenAQ API v3 | Optional nearby air-monitor observations, when the operator configures a server-side OpenAQ API key. Readings appear as separate monitoring sites and do not replace the CAMS model. | Only readings timestamped within the last 48 hours are shown. Raw pollutant values retain their units and provider context; they are not converted to AQI. Station coverage and freshness vary by region. Attribution: OpenAQ. Terms of use. |
| European Environment Agency (EEA) validated monitoring-station catalogue | European station locations and pollutant-specific year ranges for validated historical measurement files. These are shown as distinct station markers, not interpolated into an air- quality surface and not treated as current readings. | European coverage; available records and years vary by station and pollutant. Dataset license: CC BY 4.0; acknowledge the EEA. Dataset metadata and license · Official station service. |
| Vercel | Website hosting and serverless infrastructure. | Processes technical request data needed to deliver and secure the service. |
| Firebase / Google services | Not initialized by the current site build. | The current site build does not initialize Firebase Analytics or use Firebase AI Logic / Gemini for recommendations. |
| Qwen2.5-0.5B-Instruct / Transformers.js | Generates recommendations locally by default; optional hosted mode uses a serialized GeoHearth server queue. | Local mode downloads model files from Hugging Face. Hosted mode sends selected-area evidence and prompts to the configured model host. AI can make errors and must be independently checked. |
How to interpret the data
Each metric should be read together with its source, date and resolution. Missing or unavailable data should not be treated as zero. A mapped-data miss does not prove that an environmental feature does not exist. Satellite and modeled information can differ from on-site measurements.
Third-party names and trademarks belong to their respective owners. Listing a source does not imply that the provider endorses GeoHearth AI.