CARTO

The decision that defines CARTO

Spatial analysis platforms have traditionally worked by taking your data in. You export from wherever it lives, load it into the platform, analyse it there and produce maps.

That creates two problems. The data now exists in two places, with the governance and freshness questions that follow. And the platform's own capacity becomes the ceiling, which for large data sets is a real constraint.

CARTO takes the other approach. The analysis runs where the data already is.

Analysis in the warehouse

Modern cloud data warehouses have their own spatial functions. They can calculate distances, test whether a point falls inside a boundary, buffer a geometry and aggregate by area.

CARTO generates queries that use those functions. The warehouse performs the spatial work, and the platform handles expressing the analysis and presenting the results.

The consequences are direct. No copy of the data is created, so access control stays where it already is. And the compute is the warehouse's, so a spatial query over a very large table is handled by a system already sized for large tables.

Maps and widgets

Over the results, maps are built with styled layers, accompanied by widgets showing figures, distributions and filters.

Interacting with the map filters the widgets, and interacting with a widget filters the map. That linking is what makes a spatial dashboard useful rather than decorative, because a question about a region can be followed without rebuilding anything.

Workflows

Spatial analysis is often a sequence: buffer these points, join to those boundaries, aggregate by area, filter to results above a threshold.

Written as SQL that is a long query with several nested steps, and it is difficult for anyone else to follow.

Workflows express the same sequence as steps on a canvas. Each step is one operation, the order is visible, and the whole thing can be rerun when the data changes. The queries still execute in the warehouse; what changes is how the analysis is expressed and shared.

The reference data catalogue

Spatial analysis nearly always needs context. Administrative boundaries, population figures, road networks.

Sourcing that data, checking its licensing and loading it correctly is a project of its own, and it is where spatial work often stalls before it starts.

A catalogue of reference data addresses this by making those sets available to join against directly.

Who uses CARTO

CARTO is used where location is central to a commercial question: retail site selection and catchment analysis, telecommunications and utility network planning, logistics and delivery operations, and public sector analysis by area.

Its users are analysts working with spatial questions, in organisations that already have a cloud data warehouse.

Points to consider

A supported cloud data warehouse is a prerequisite. The approach depends on one, so it is not a candidate where data sits in files or a conventional database.

Query cost also falls on the warehouse. Spatial functions over large tables consume credits, which is the same trade any warehouse based tool involves and is worth understanding before opening it to a wide group.

Getting started

The documentation covers connecting a warehouse, building a first map from a table, using widgets and building a workflow, along with the catalogue of reference data. Connecting a table that already has coordinates and mapping it is the usual introduction.