Skip to main content
Whether you’re new to Wherobots or to geospatial data, this glossary provides short definitions for the terms you’ll encounter throughout these docs, with a link to the page that covers each one.

​
Wherobots Platform

Core terminology for the Wherobots platform, its products, and key concepts.

​
Products

Wherobots has two products, WherobotsDB and RasterFlow, which you use in Wherobots Cloud. GeoStats and map matching are WherobotsDB modules.
WherobotsDB is the Wherobots lakehouse compute engine for spatial data. It runs Spatial SQL, Python, and Scala workloads on Apache Iceberg tables and files in object storage, with 300+ vector and raster functions and Apache Sedona API compatibility. GeoStats, map matching, and vector tile generation are WherobotsDB modules.WherobotsDB in depth
RasterFlow is the Wherobots AI perception engine for Earth imagery. It builds mosaics from satellite and aerial imagery, runs computer vision models over them at scale, and vectorizes the predictions for analysis in WherobotsDB.Get started with RasterFlow
WherobotsDB modules: GeoStats runs spatial statistics such as DBSCAN, Getis-Ord Gi*, and Local Outlier Factor, and map matching aligns GPS traces with road networks.GeoStats and map matching overview

​
Tools and interfaces

The managed service and web console where you use WherobotsDB and RasterFlow. In it, you start runtimes for notebooks, Job Runs, and SQL Sessions, connect data, and manage your Organization.Wherobots Cloud
A command-line interface for submitting and monitoring Job Runs, retrieving logs, and calling the Wherobots API from your terminal or from agentic coding tools.Install the Wherobots CLI | Agentic terminal usage
A Model Context Protocol server for exploring spatial data catalogs and generating Spatial SQL queries through natural language in AI-powered IDEs.Set up the MCP Server
An extension for VS Code and compatible Code OSS editors that manages workspaces, connects local Jupyter notebooks to Wherobots runtimes, and submits Job Runs.Install the Wherobots VS Code Extension
An API, with Python and Java client SDKs, for running Spatial SQL queries against WherobotsDB without a notebook. Connections use a SQL Session, which can be created or reused.Spatial SQL API docs

​
Deployment Models

Every workload runs in a region, which is either Hosted or BYOC.
The default region type, in which your workloads run on serverless compute in the Wherobots AWS account, the counterpart to a BYOC region.Hosted and BYOC regions
A region type in which Wherobots runs your workloads on infrastructure it manages inside a VPC in your own AWS account, instead of on the serverless compute of a Hosted region. BYOC is available to Innovation and Enterprise organizations.Learn about BYOC

​
Compute and Billing

The credit card, bank account, or AWS Marketplace subscription your Organization pays with, which keeps your Organization on the Professional Edition after the free trial. A payment method is what you add; a billing method is how it’s billed.Add a payment method
How your Organization is billed for its usage: Direct Billing by Wherobots, through the AWS Marketplace, or through a contract with Wherobots.Change your billing method
A billing method in which you pay for the Professional Edition with a credit card or bank account directly in Wherobots Cloud, without an AWS Marketplace subscription.Change your billing method
The amount of usage your free Professional Edition trial covers. The trial ends when the trial period elapses or the allowance is used up, whichever comes first.How the free trial works
The period of inactivity, 15, 45, or 120 minutes, after which a notebook runtime scales down. You are still charged for minimum resources until the notebook is destroyed.Idle timeout for Wherobots Notebooks
Your Organization’s maximum concurrent computing power, measured in Spatial Units, which depends on your Organization Edition and payment history.Runtimes and quota
The unit that measures RasterFlow usage, based on the data a task processes (pixels × bands × time periods) and the task’s complexity factor. Unlike a WherobotsDB Spatial Unit, it doesn’t depend on how long a runtime runs, so you can estimate a task’s cost before you run it.Understand RasterFlow billing
A single unit of RasterFlow work that is metered and billed on its own: mosaicking, model inference, Build Multiscales, or Mosaic Vectorization. Each task consumes RasterFlow Spatial Units at its own price.Understand RasterFlow billing
A dedicated, serverless compute cluster that runs your Wherobots workloads. Runtimes come in two types, General Purpose and Memory Optimized, and in sizes from Micro to 4X-Large.Learn more about runtimes
The maximum duration a runtime instance can remain active, 8 hours by default. You can request an increase through a Compute Request.Runtime time-to-live
The billing unit for the compute provisioned to a runtime, also called a WherobotsDB Spatial Unit to distinguish it from a RasterFlow Spatial Unit. Cost is Spatial Units multiplied by the hourly price.Understand Spatial Unit pricing
Any computational task that runs on a Wherobots runtime, including Job Runs, SQL Sessions, and notebooks.Workload History
A session in the Wherobots VS Code Extension that provisions a Runtime for your local Jupyter notebooks to run against. Stopping a workspace tears down the runtime and halts Spatial Unit consumption.Manage workspaces

​
Data and Storage

Querying data across multiple catalogs and storage systems, such as your own S3 storage, Managed Storage, and Unity Catalog, from a single Wherobots runtime without duplicating data.Learn about Data Federation
The Wherobots Cloud interface where you browse and create catalogs, configure storage integrations, and find the datasets Wherobots provides.Configure Data Hub
Havasu: the former name of the Wherobots table format. Wherobots tables are Apache Iceberg tables; the name survives in identifiers such as havasu.iceberg and havasu.* table properties.Upgrade existing geometry tables
Cloud storage that Wherobots provisions and manages for your Organization, ready for reading and writing data without configuring a storage integration.Managed Storage
A connection between Wherobots and your own Amazon S3 buckets that lets your runtimes read and write data directly in your cloud storage.Set up S3 integration
The entry point for WherobotsDB code. Creating a SedonaContext initializes WherobotsDB and the underlying Spark cluster, and it is the first step in any Wherobots notebook or job.Set Up WherobotsDB
A Databricks data governance layer that Wherobots connects to as a foreign catalog, so you can query Databricks-managed tables without moving data.Connect to Unity Catalog
The wherobots_open_data catalog, which Wherobots manages, of public geospatial datasets that are ready to query in any Wherobots runtime.Introduction to the Data Hub

​
Organization and Access

A scheduled or on-demand execution of a workload outside the notebook environment, which you can orchestrate from Apache Airflow with the WherobotsRunOperator.Learn about Job Runs
A Workload type that starts a WherobotsDB runtime for running Spatial SQL queries programmatically, used by the Spatial SQL API, the Wherobots MCP Server, and client SDKs. Each session is billed in Spatial Units for its active duration.Spatial SQL API docs
The three tiers of Wherobots service: Professional, Innovation, and Enterprise. Each edition adds features, capacity, and support to the one before it.Compare editions
A non-human identity for programmatic access to the Wherobots API, used for automation, CI/CD pipelines, and integrations without tying access to a user account.Set up Service Principals

​
General Geospatial Terms

Foundational geospatial concepts and standards referenced throughout the Wherobots documentation.

​
Spatial Data Types

Spatial data stored as a grid of cells (pixels), where each cell holds one or more values such as elevation, temperature, or reflectance. Common formats include GeoTIFF / COG, netCDF, and Zarr.Introduction to Spatial Data
Spatial data that represents discrete features as points, lines, polygons, and collections of them. WherobotsDB stores vector data with the Geometry or Geography type.Introduction to Spatial Data
The “round Earth” spatial data type. A geography column interprets coordinates on a sphere, so ST_Distance and ST_Length return spherical results in meters, and ST_Area in square meters, without projecting first.Geography columns
The planar spatial data type, which treats coordinates as Cartesian (x, y) values and computes distances and areas on a flat plane. Geometries follow the OGC Simple Features specification.Geometry Functions

​
Spatial Formats

A GeoTIFF file with internal tiling and overviews, so clients can read only the portion they need over HTTP range requests.Raster loader
A JSON-based format for geographic features, widely used in web mapping and APIs. It is human-readable but less efficient than binary formats for large datasets.Load GeoJSON data
An OGC standard built on SQLite for storing vector features, tile matrices, raster data, and metadata in a single portable file.Load from GeoPackage
An extension of Apache Parquet that adds standardized geospatial metadata and geometry encoding. WherobotsDB has native read and write support for GeoParquet.Load GeoParquet
A single-file archive format for vector or raster map tiles that can be served directly from cloud object storage over HTTP range requests, without a tile server.Tile Generation
A widely supported vector format from Esri, stored as a collection of files (.shp, .shx, .dbf, and others). Its limits include a 2 GB file size cap, 10-character field names, and no built-in UTF-8 support.Load Shapefile
The binary equivalent of WKT, a compact, machine-readable encoding of geometry objects.ST_GeomFromWKB
A text format for geometry objects and coordinate reference systems, such as POINT(30 10) or POLYGON((0 0, 1 0, 1 1, 0 1, 0 0)).ST_GeomFromWKT
A format for chunked, compressed, N-dimensional arrays. RasterFlow stores mosaics, model outputs, and multiscale stores in Zarr.Visualize RasterFlow Outputs

​
Spatial Concepts

A geographic boundary that scopes an analysis or data request. In RasterFlow, you pass the AOI as a WKT string to set where mosaics are built and inference runs.RasterFlow Datasets
A framework that defines how coordinates map to locations on Earth. A geographic CRS such as WGS 84 uses latitude and longitude, and a projected CRS such as Web Mercator (EPSG:3857) flattens the curved surface onto a plane.CRS Transformation
Converting an address or place name into geographic coordinates. Reverse geocoding converts coordinates into an address or place name.ST_ReverseGeocode
A virtual boundary around a real-world area that triggers an event when a tracked device, vehicle, or person enters or exits it. Geofencing relies on spatial predicates such as ST_Contains or ST_Intersects.ST_Contains
An algorithm that aligns noisy location data, such as GPS traces, to the most probable path on a road network. In Wherobots, map matching is a WherobotsDB module.Map Matching tutorial
Ingesting raw geospatial data, applying spatial transformations such as reprojection, clipping, enrichment, and joins, and loading the results into a target system.Data engineering
A data structure or data layout, such as an R-tree, a quadtree, or a Hilbert curve ordering, that speeds up spatial queries. WherobotsDB optimizes spatial joins automatically, and the CREATE SPATIAL INDEX statement rewrites a table’s data files clustered by a geometry column so spatial filters can skip files that don’t match.Cluster by Geospatial Fields for Faster Queries | Optimizers
A join where the condition is a spatial relationship, such as ST_Intersects, ST_Contains, or ST_DWithin, rather than a key match. For example, joining building footprints with parcel boundaries finds the buildings in each parcel.Optimizers
The World Geodetic System 1984, the geographic CRS with latitude and longitude in degrees that GPS, web mapping, and most interchange formats use. EPSG:4326 is its identifier in the EPSG registry.CRS Transformation
Summarizing raster cell values within vector zones (polygons), such as the average elevation or total rainfall within each county boundary.RS_ZonalStats

​
Remote Sensing and Raster Analysis

A raster dataset of bare-earth surface elevation, used for terrain analysis, flood modeling, and slope mapping. Digital Surface Models (DSMs) also include buildings and vegetation.Digital elevation models
Running a trained computer vision model over raster data to produce per-pixel predictions, such as semantic segmentation classes, object detection boxes, or canopy height. In RasterFlow, inference runs over a mosaic, writes a Zarr store, and is billed as a RasterFlow Task.RasterFlow Models
A composite raster stitched together from multiple overlapping images. In RasterFlow, a mosaic is a spatially aligned raster dataset in Zarr format that covers an Area of Interest and time range.RasterFlow Datasets
Building a mosaic by reading source imagery across an Area of Interest, reprojecting it onto a common output grid, and combining overlapping observations. In RasterFlow, mosaicking is a billable RasterFlow Task.RasterFlow Datasets
Downsampled overview levels, an image pyramid, added to a Zarr store so a map client can view it interactively. Building multiscales writes a new store and leaves the source store unchanged.Build Multiscales with RasterFlow
A spectral index of vegetation health computed from near-infrared and red reflectance. NDVI values range from -1 to 1, where higher values indicate denser vegetation.Raster map algebra
A computer vision task that locates discrete objects in an image with bounding boxes. In RasterFlow, object detection finds features such as ships, solar panels, and aircraft in satellite imagery.RasterFlow Models
Collecting information about the Earth’s surface from sensors on satellites or aircraft. Remote sensing produces the raster imagery used for land cover classification, change detection, and environmental monitoring.RasterFlow Datasets
A computer vision task that classifies every pixel in an image into a category, such as water, forest, building, or road.RasterFlow Models
Converting raster data into vector geometries. In RasterFlow, vectorization turns pixel-level model predictions into polygons stored in GeoParquet.Get Started with RasterFlow

​
Ecosystem and Standards

An open-source platform for authoring, scheduling, and monitoring data pipelines. Wherobots integrates with Airflow through the WherobotsRunOperator to schedule Job Runs.Airflow configuration
An open table format for large analytic datasets, with ACID transactions, schema evolution, time travel, and partition evolution. Wherobots tables are Iceberg tables; see Havasu for the former name.Geometry Support
An open-source cluster computing system that extends Apache Spark and Apache Flink with distributed spatial datasets and Spatial SQL. Wherobots was founded by the creators of Apache Sedona, and WherobotsDB is compatible with its APIs.Apache Sedona vs. Wherobots
A unified analytics engine for large-scale data processing that WherobotsDB extends with geospatial capabilities. Creating a SedonaContext initializes a Spark cluster under the hood.Set Up WherobotsDB
An open-source translator library for 200+ raster and vector geospatial data formats, and a building block of most geospatial software.Raster Data in WherobotsDB
An international organization that develops open standards for geospatial content and services, including Simple Features, GeoPackage, and WKT/WKB encoding.
An open data initiative that produces map datasets covering buildings, places, transportation networks, and administrative boundaries. Wherobots hosts Overture Maps data, with historical version snapshots, in the Wherobots Open Data catalog.Explore Overture Maps data
SQL extended with spatial data types and functions, such as ST_Intersects, ST_Buffer, and ST_Area, following the OGC/ISO SQL/MM spatial standard. Spatial SQL is the primary query language in WherobotsDB.Write Spatial SQL queries
A specification for cataloging spatiotemporal data, primarily satellite imagery, in a standardized, searchable way.STAC data source
An open-source Python library maintained by Wherobots that returns administrative boundaries from Overture Maps Foundation data, using dot-notation such as wkls.us.ca.sanfrancisco.Find geographic boundaries with wkls