Data Sources

Elements analyzes data from Data Sources. Data Source can be EO imagery, SAR imagery, geolocation data and even your own custom data. A data source consists of one or more observations, which generally have a geospatial and time component, plus additional attributes.

Installed Data Sources can be reviewed using the Data Explorer. The Data Explorer lets users interactively search for data based on a variety of filters including:

• Data Source(s)
• Location
• Time
• Data Source attributes

Standard APIs

Elements exposes Data Sources via industry-standard APIs to make it easier to integrate with your existing applications. These APIs include:

• STAC
• Queryables
• Search and filter language (CQL2)
• Schema
• Features

STAC

STAC, which stands for spatial temporal asset catalog, is a grass-roots driven specification for cataloging imagery data. It is a REST based API that returns results using GeoJson, and more recently, GeoParquet. It is seeing growing adoption in the industry.

Queryables

The queryables API is based on the STAC filter queryables extension. It is advertised via a JSON Schema document retrieved from the /queryables endpoint. The root endpoint retrieves queryables that apply to all collections. Additional endpoints provide queryables for subcollections such as Planet Skysat data. For imagery data, the queryables API should return the list of metadata fields specified by the image provider. For geolocation data, the queryables API should return a list of all searchable columns available in the data source. The queryables endpoint returns a JSON schema of the Data Source fields that can be used to create filters. It is based on the following two specifications:

Search

The search API returns items from a collection that match a specified filter. At its simplest the filter can be a date or bounding box, but more sophisticated filters can be created using the CQL2 query language.

Once again the specifications come from STAC and OGC.

The Search endpoint differs between imagery data sources and tabular data sources such as geolocation. In the case of imagery data, the search request goes to the STAC endpoint and returns GeoJson. For vector data, the search request goes to the Feature API and returns geoparquet files.

The search API finds data based on search filters created from the Queryables API. The search API is implemented using STAC and the STAC filter extension. Filters use CQL2 which is an OGC standard that supports spatial and temporal searches in addition to standard searches.

The search api endpoint is /search and returns a list of items. Each item in turn may have multiple assets. Results are returned in GeoJson. These results can be visualized by clients - for example the following screenshot shows Planet Skysat images found for a specific data range and location.

Schema

The schema API returns a JSON schema that returns the properties in a data source. For example, an AIS dataset would include an id column of type string, a longitude column of type double, etc. These may differ from the list returned by the queryables endpoint which is focused on properties that can be queried.

The schema language is JSON schema and is defined by OGC Features specification in https://portal.ogc.org/files/108199.

Features

For geolocation data, items are individual features such as a single ping from a ship. Behind the scenes, the items endpoint is implemented via an OGC features items (see https://docs.ogc.org/is/17-069r4/17-069r4.html#_features_rootcollectionscollectioniditems).