Microsoft’s Planetary Computer vs. Google Earth Engine: A Compare and Contrast

Traditional geospatial analysis involves working with locally hosted raster and vector data that has been downloaded from a source and stored on the hard drive of the processing machine. As the needs and capabilities of technology expand, it has become necessary to outsource both data storage and computation to the cloud, allowing users to work with larger datasets and save on time and processing power. Multiple advanced platforms are emerging to meet this need.

Google Earth Engine (GEE) was launched in 2010, making it one of the first cloud-based platforms of its kind for large-scale geospatial analysis. Not only did it host massive amounts of satellite imagery, but it also enabled users to leverage the Google cloud to conduct processing and analysis on that data. This eliminates the need for local computing power or specific software on the part of the user.

In April 2021, Microsoft launched its Planetary Computer (PC) in public preview, citing sustainability and climate change research as its motivation for making petabytes of earth observation data available online. Like Google Earth Engine, Microsoft’s Planetary Computer is not just a storage mechanism; equipped with its own API and development environment, PC is a powerful tool for computation and geospatial analytics.

Google Earth Engine is free for non-commercial use, including academic research, education, and nonprofit activities. Users register with a Google account. Commercial and government operational use is subject to subscription and compute charges. Google Earth Engine does favour academic use — registration with a “.edu” email tends to be approved immediately, whereas other users may take longer to be approved. This means students will likely be entering the workforce with GEE experience baked into their education; coupled with its long head start in the market, this appears to give GEE a distinct advantage. On the other hand, the Microsoft suite is considerably entrenched in the corporate space, and PC integrates seamlessly into the workflows of users of its existing cloud computing service, Azure.

Planetary Computer is actively integrating open-source technology, which gives this service interoperability with many other tools that your organisation may already be using. Microsoft wants to enable users to access the best-in-class open-source tools, as well as develop new additions.

“I believe [the geospatial open-source ecosystem] is such a valuable asset, not just in the fight for climate change, but in general,” says Rob Emanuele, a geospatial architect at Microsoft. “Anything that we’re building should try to take advantage of that as much as possible and contribute back to that as much as possible.”

The data available on these platforms is cloud-optimized and streamable. Essentially, at any given time, the service is only reading the packet of data that the user needs in that moment, which saves on computing power. The raster data available on both platforms includes everything from climate and weather data to terrain and elevation layers, biodiversity and biomass models, and of course, satellite imagery.

In addition to providing the raw Sentinel and Landsat imagery, Google Earth Engine provides atmospherically corrected and cloud-masked versions of the data as well. Google has also released Dynamic World, a 10 m near-real-time (NRT) Land Use/Land Cover (LULC) dataset developed in partnership with the World Resources Institute, which is available through GEE and produces updated global predictions every two to five days based on Sentinel-2 imagery.

Microsoft PC relies on the STAC (SpatioTemporal Asset Catalog) ecosystem for spatiotemporal searchability, which includes encoding metadata with a common JSON schema to specify what the data contains. This indexing mechanism helps users find exactly what data they need for analysis. You can think of it as a massive normalised database. Meanwhile, Google has a unique hierarchical nomenclature system, which makes finding specific datasets straightforward.

In terms of data processing and analysis, both services have a degree of learning curve. In Google Earth Engine, the user manipulates data using the JavaScript, Python, or REST APIs. The Planetary Computer provides its own API conforming to the STAC API specification, and Microsoft previously offered a JupyterHub-based computational environment for interactive analysis; however, that Hub was retired in June 2024. Users can connect to the Planetary Computer data and APIs from their own computational environments, including notebooks hosted on Azure or other cloud platforms.

Google Earth Engine’s catalog has grown substantially since its launch and now contains over 90 petabytes of analysis-ready geospatial data, including more than 1,000 curated datasets. The Planetary Computer data catalog has grown to over 50 petabytes of environmental monitoring data across more than 120 datasets. Google Earth Engine boasts near-instantaneous results to queries and processing tasks. GEE has historically been strongest for pixel-based raster analysis, though it also supports vector operations. The Planetary Computer’s reliance on open-source standards such as STAC and cloud-optimized formats provides strong interoperability with tools such as GDAL, Rasterio, and Dask.

Explore some of the other similarities and differences between Google Earth Engine and Microsoft Planetary Computer in this summary table.

Ultimately, choosing between the two will come down to preference and need. GEE is a more established service with a large community and extensive documentation, making it well suited for academic and research use. Microsoft’s Planetary Computer appeals to organisations already invested in the Azure ecosystem and those working primarily with open-source geospatial tooling. Both platforms have substantial data catalogs, ample cloud storage, and significant processing power.

If you are looking for more information on Google Earth Engine, listen to our podcast episode here with Qiusheng Wu.

To learn more about the Planetary Computer, try this podcast discussion with Rob Emanuele from Microsoft.

Recommended Podcast Episodes