2024 GDAL user survey
In October 2024, The GDAL maintenance program created an open survey to collect feedback on user's experience with GDAL and the direction of the maintenance program. The survey was publicized on gdal.org, the gdal-dev mailing list, the project GitHub page, and social media. From October 28 to November 21, the survey received 602 responses.
Who responded to the survey?
Survey respondents were generally very experienced users, with 79% of users having spent 5 or more years working with GDAL. Surprisingly, two respondents claimed to be Frank Warmerdam, who originated the project in 1998. More than half (52%) have built GDAL from source, 29% subscribe to the gdal-dev mailing list, and 29% have contributed to the project by submitting bug reports or pull requests. The high experience level of respondents reflects the challenge of reaching users who may use GDAL less often, through other software, or are not connected to the project community via mailing lists or social media.
Operating system
Most survey respondents use GDAL on Linux, followed by Windows and OS X. Responses to "Other" included WSL2 and iOS.
Local or cloud?
Most survey respondents use GDAL primarily with local file systems.
Data formats
Among raster data formats, GeoTIFF commands an overwhelming majority of GDAL usage:
The most popular vector format was GeoPackage, followed by classics such as Shapefile, GeoJSON, and PostGIS. GeoParquet, Esri FileGeodatabase, FlatGeobuf, and GML each earned enough votes to remain out of the "Other" category.
Installing GDAL
Survey respondents obtain GDAL from a variety of channels, depending on the platform. On Linux, standard system packages are the most popular solution. OSX users rely primarily on Homebrew; most Windows users user OSGeo4W. The popularity of Homebrew among Windows users may indicate that GDAL is being used through the Windows Subsystem for Linux (WSL). The reported usage of OSGeo4W by Linux users is more difficult to explain.
As may be expected for a group of experienced users, most respondents reported that GDAL is easy to install with the options they need. Still, installation remains a difficulty for many users.
Installation difficulties were not associated with a particular operating system.
How is GDAL used?
The greatest number of respondents reported using GDAL from Python, with a roughly 50/50 split between the GDAL Python bindings and higher-level packages such as shapely, rasterio, and geopandas. After Python, the greatest number of respondents reported using the command line interface, followed by smaller number of users working in R, PostGIS, and QGIS.
Getting help with GDAL
Most users use gdal.org (directly or via a search engine) as their starting point when trying to get help with GDAL.
Difficulties using GDAL
Users did not identify a single area as a source of their challenges with GDAL. However, the top responses of "finding examples" and "understanding features" point to a shortage of documentation.
Consistent with the above, respondents reported "examples", "workflows", and "API usage" as high priorities for documentation efforts.
And "examples" and "doc" rank highly among open-ended responses to "what could make GDAL easier to use?"
Maintenance program activities
Among activities undertaken by the maintenance program so far, respondents found the most value in enhancements to GDAL's dependencies (such as PROJ, GEOS, and libtiff), its Python bindings, and documentation.
Asked about a variety of tasks the maintenance program could take on beyond those listed above, respondents showed some enthusiasm for almost everything! Still, high priorities were given to performance, improving format capabilities, and improving the command line interface while preserving backward compatibility.
Next steps
The maintenance program will use these results to inform work over the coming year. Some work has already been performed to develop an improved command-line interface and add a mechanism for usage examples to be cross-referenced in the documentation.