Data sources and attribution

Maps and geographic context

The interactive basemap uses OpenFreeMap with data attributed to OpenStreetMap contributors and the map style’s attribution. OpenStreetMap data is available under the Open Database License. The map retains its attribution control. A visible building footprint is context, not a verified building height.

The world day/night map uses Natural Earth public-domain map data. Its small-scale coastlines are for world orientation, not property boundaries. The terminator and subsolar point are calculated, not satellite photographs.

Optional imagery, terrain and buildings

Satellite mode uses EOX::Maps and EOxCloudless by EOX IT Services GmbH, containing modified Copernicus Sentinel data 2017, under CC BY 4.0. The 2017 mosaic has 10 m source pixels; it is historical regional context, not live or survey imagery. Availability and rate limits depend on EOX.

Terrain import uses Mapzen Terrain Tiles hosted on AWS. Attribution: Mapzen, USGS, 3DEP, SRTM, GMTED2010, ETOPO1 and other sources. Source resolution, acquisition dates and accuracy vary. SunCarta decodes Terrarium elevations at zoom 12, samples nine distances from 100 m to 12 km along 72 bearings, includes geometric Earth curvature, and interpolates between horizon bearings. Narrow features can be missed. The result affects modeled direct sun, not standard astronomical sunrise or alignment windows.

Nearby building import queries Overpass API for OpenStreetMap ways within 300 m, capped at 300 ways. Only tagged heights or explicitly labeled 3 m/storey estimates become simplified vertical screens. Relations, enclosing footprints, invalid geometry and missing heights are skipped. The imported data remains attributed to OpenStreetMap contributors under ODbL. Loading is optional, and the default scene remains empty. No tree-canopy coverage is inferred.

Solar position and independent references

Astronomy Engine 2.1.19 supplies the locally calculated ephemeris under the MIT license. Independent validation uses airless topocentric JPL Horizons positions and USNO event data. These are comparison references, not a live API dependency for slider movement.

The methodology defines solar-centre elevation, true-north azimuth, atmospheric event conventions, supported dates and the measured limits of the fixture comparisons.

Time zones and search

Offline coordinate lookup selects an IANA time-zone identifier. JavaScript’s runtime time-zone database resolves civil rules at the actual instant. Historical rules and future government changes can differ between runtime database versions.

Address search calls SunCarta’s same-origin /api/geocode endpoint, which validates queries and sanitizes Photon-compatible results. The public Photon service uses OpenStreetMap-derived place data; provider availability and quotas can limit search. Coordinate entry, browser geolocation and map placement are alternative inputs. A geocoded address does not identify the exact window or garden bed.

Scene data and illustrations

Obstacle screens, receiver height, manual horizon angles and window dimensions are visitor inputs. SunCarta does not verify them or silently import global building heights, tree canopies or terrain. Scenes stay in local browser storage by default; chosen exports and share links include their coordinates and modeled geometry.

The interior editorial illustration is an original generated conceptual image. It is not a NASA photograph, a measured room, or evidence of how much light enters a property. The interactive room itself is a schematic geometric projection.

Coverage and future data

Footprints, building heights, bare-earth elevation, surface elevation and vegetation are different datasets. A footprint cannot substitute for a height; a terrain model cannot substitute for tree crowns. Provider availability at a city level does not establish complete coverage around a chosen point.

The release’s 24-place research pilot includes incomplete provider responses and unverified heights. Automatic global shadow coverage is therefore not claimed. Missing data remains unknown. Use a measured screen or manual horizon where appropriate, and keep unmodeled features in your notes.