The GeoAnalytics Server toolbox contains a set of powerful tools for performing spatial analysis on big data. GeoAnalytics Tools are powered by your ArcGIS GeoAnalytics Server. ArcGIS GeoAnalytics Server distributes the analysis between multiple server nodes. By using distributed processing, you can process large datasets in less time.
GeoAnalytics Server tools are designed for large datasets; consequently, other desktop or Enterprise tools may be more appropriate for use with smaller datasets. GeoAnalytics Server tools require an initial startup time to implement the distributed processing, so they are optimal for larger datasets (hundreds of thousands or millions of records).
Similar to other tools in ArcGIS Pro, the performance of GeoAnalytics Server tools depends on the following:
- The size of the input data—Such as the number of features and number of fields.
- The input data source—For example, when stored on the same machine, feature services are slower than big data file shares.
- The tool you are running—For example, Aggregate Points completes execution quicker than Calculate Density with the same data and bin size.
- The parameters you use in the tool—For example, when using the Join Features tool, a smaller join distance performs better than a larger one.
- The hardware on your GeoAnalytics Server machines.
Each GeoAnalytics Server tool topic includes a usage note about improving tool performance by modifying parameters.
Data
When running analysis, data that is colocated with your GeoAnalytics Server will perform best.
Inputs
GeoAnalytics Server tools support the following data sources for input:
- Feature layers and feature services (including hosted)
- Big Data File Shares
- Data local to ArcGIS Pro including shapefiles and file geodatabases.
Using local data such as shapefiles and file geodatabases is not recommended for most datasets. Instead of using local data, share your layer as a hosted feature layer or create a big data file share. Additionally, using data from a different Enterprise deployment than your deployment with GeoAnalytics Server behaves similarly to local data and may also be slow. If you must use large services from another deployment, it is recommend that you run the tool through Enterprise Map Viewer Classic or ArcGIS REST API.
GeoAnalytics Server tools do not support the following data sources for input:
- Geopackages
- XY event layers
If you are using big data file share inputs, you cannot visualize your inputs on the map, but you can browse and analyze them in the GeoAnalytics tools. If you need to apply a filter to the inputs, it is recommended to your Enterprise Map Viewer Classic, which is faster than copying the layer to an ArcGIS Data Store and using the resulting feature service as an input. Generally, it is not recommend that you spend time moving your big datasets around. Instead, use the flexibility of big data file shares as a way to use your data in its current format.
Outputs
If you want to visualize your results or maintain them in your Enterprise Contents, save your results as a hosted feature layer. This is the default. Results stored as a hosted feature layer can be stored in the relational data store or the spatiotemporal data store—this is the default. Results stored in the spatiotemporal big data store are stored in WGS84. If you want to store your results in a different spatial reference, save your results to the relational data store.
If you want to save your data to a big data file share for further analysis or to be used elsewhere in your organization, save it as a big data file share dataset by setting up a template. Big data file share outputs cannot be visualized on a map. Saving results to a big data file share is available at ArcGIS Enterprise 10.7 or later and cannot be completed with ArcGIS Pro. Use Enterprise Map Viewer Classic or ArcGIS REST API to save results to a big data file share.
Analysis
GeoAnalytics Server tools are implemented differently to other ArcGIS Pro tools.
GeoAnalytics does not include polygonal slivers in its operations. In the Join Features and Overlay Layers GeoAnalytics tools, slivers are not included in the analysis, and some very small features may be excluded from results.
GeoAnalytics Server tools result in less densified features than other ArcGIS Pro tools. For example, the following images depict the result of vertices on a buffered polygon using Buffer in the Analysis toolbox (blue), and Create Buffers in the GeoAnalytics Desktop toolbox (orange). The image on the left shows the buffered polygons overlayed on each other, and the image on the right shows a zoomed-in view of some of the polygon vertices. 
GeoAnalytics Server tools do not support the in_memory or memory workspace.
When running GeoAnalytics Server tools, the analysis takes place in memory on the GeoAnalytics Server machines. When the data being analyzed doesn't fit in memory, it is written to disk in your specified directory. If your tool isn't completing and is running out of space, you can modify which directory to use.
Best practices
When running analysis, it is best to analyze only the data you are interested in. You can limit the data you analyze by doing the following:
- Apply a definition query to a layer on your map. This is not an option for big data file share inputs.
- Apply a selection to features on your map. This is not an option for big data file share inputs.
- Set the processing extent of your analysis to limit the spatial extent of features used.
- Use the time slider to specify the extent of data to analyze. This is only applied to layers that are drawn on the map.
When using GeoAnalytics Server tools, a definition query typically processes more quickly than a selection. Currently, you cannot apply a filter to big data file share layers in GeoAnalytics Server. Use Enterprise Map Viewer Classic or ArcGIS REST API to apply a filter to big data file share layers.
Use time in analysis
Many GeoAnalytics Server tools use or require time. To take advantage of time stepping, temporal joins, or track-based analysis (for example, using Reconstruct Tracks, Find Dwell Locations, or Detect Incidents), your layers must be time enabled. You can enable time on the following datasets:
- Feature layers—Enable time when sharing your layer or editing the item.
- Big data file shares—Edit the big data file share manifest to enable time.
When you use GeoAnalytics Server tools in ArcGIS Pro, it does not use the time zone or daylight savings options specified in the time settings. Instead, it does the following:
- Feature Services—Layers in your portal contents are in UTC. You cannot modify the time settings.
- Local layers—Layers such as a local feature class or shapefile are converted to UTC time. This is done by calculating the time zone your GeoAnalytics Server machine resides in. For example, if your GeoAnalytics Server is in Pacific time (-8:00 hours), and you have a feature that has a time stamp of 1 p.m. when you run analysis, that feature will be at 9 p.m. It is recommend that you share layers to your portal to have more control over the time zones. Daylight savings is always applied.
- Big data file shares—These are not time-enabled in ArcGIS Pro. Instead, when you define the manifest, you specify the time zone that will be used.
Layers added to the map (such as feature services and local data) only analyze features in the visible time extent.