With the emergence of multi-core GPUs, new database technologies have been developed to take advantage of this improved technology. MapD, a startup based in San Francisco, is one example of these companies. Their GPU-based database technology was made open source in 2017 and is available for use on cloud services, such as Amazon Web Services (AWS) and Microsoft Azure. By combining the parallelization potential of GPUs with a relational database, the MapD database improves the speed of database queries and visualizations based on the data. 

MapD has created a Python 3 module, pymapd, that allows users to connect to the database and automate queries. This Python binding allows geospatial professionals to integrate the speed of a GPU database into an existing geospatial architecture, adding speed improvements to analysis and queries. Both of MapD's core offerings (the open source community version and the commercial enterprise version) are supported by pymapd. 

In addition to the Python module, MapD has added geospatial capabilities to their database technology. Storage of points, lines, and polygons is now supported, as is a spatial analysis engine that offers distance and contains functionality. Also, MapD has developed a visualization component, Immerse, that allows for analytical dashboards to be built quickly, with the database as a backend. 

In this chapter, we will cover the following topics:

• Create a GPU database in the cloud
• Explore data visualizations using Immerse and the SQL EDITOR 
• Use pymapd to load spatial and tabular data into the database
• Use pymapd to query the database
• Integrate the cloud database into a GIS architecture