What's new—ArcGIS Velocity

The June 2021 update of ArcGIS Velocity includes enhancements and new functionality. Highlights are provided below.

General

The model view analytic editing experience has been rebuilt for improved performance and includes new enhancements such as easier node resizing, node renaming, and smart snapping during manual model arrangement. Additionally, the links between analytic steps use animation to reflect if the analytic is currently running:
- If the lines are animated, then the current view of the analytic reflects what is actively running.
- As an analytic is edited such as adding additional inputs, analytic steps, or outputs, the animation will stop as the current view of the analytic is in draft mode and no longer reflects what is actively running. As before, when an analytic is saved, if it is actively running it will restart in order to pick up the changes.
Improvements were made to the entry of raw sample data when configuring a feed or data source. When samples cannot be obtained to support deriving the data schema, you have the option to provide your own sample data. The user experience was enhanced to provide better indication of when the input sample data has incorrect JSON formatting.
A variety of enhancements were made to the Velocity application to improve performance such as eliminating redundant item status requests.

A new feed type has been added to ingest vehicle positions from Geotab devices. Geotab provides fleet management software and GPS tracking devices and is used by many organizations to monitor their public works, logistics, and other types of vehicle fleets. For details, see the Geotab feed type.
The HTTP Poller feed type was enhanced for additional OAuth 2.0 security paradigms such as enabling users to manually specify the configuration even if the destination API does not have appropriate CORS headers and allow apps to make queries as well as buffering the refresh of a token by several seconds to avoid unexpected token expiration.
The Azure Cosmos DB data source was enhanced with support for the analytic runtime variables within the definition of the WHERE clause. This supports creating big data analytics that retrieve the newest data from an Azure Cosmos DB data source on every run. Additionally, improvements were made for the handling of ISO8601 strings designated as date fields.
The Text Message output was enhanced to support international phone numbers.
XML format support was enhanced to align with JSON parsing support and include all the nested fields below a root object, and GeoJSON was enhanced with improved handling for empty objects in sampled data.

Dynamic geofencing support is now available for beta testing in real-time analytics. Dynamic geofencing enables two real-time feeds to be used together to find spatial relationships, perform spatial enrichment, and make decisions based on spatial proximity. In this release, the tools that can leverage dynamic geofencing in beta mode and be used with two feeds are:
Note:
As beta functionality, real-time analytics configured with dynamic geofencing should not be used in production workflows. When a real-time analytic is configured with any of the above tools and uses two feeds as inputs, validation will report a warning that the analytic is leveraging beta capabilities. For more information, see Geofencing analysis.
A new tool, Forest-based Classification and Regression, was added for big data analytics. This tool can be used to create models and generate predictions via supervised machine learning. Predictions can be performed for both categorical and continuous variables.
Snap to Network was enhanced with an option to retain all features, including those that could not be successfully snapped to the given road network. Additionally, improvements were made to handling of road network features that do not have the required from / to information on a road segment.
Improved logging for cases where an output layer cannot be published as defined because the name is already used by another layer in the organization.

Data retention is now supported when storing output data and keeping the latest observation for any given Track ID. For more information, see the Feature Layer (new) output type and About data retention.
Enhancements were made to eliminate layer downtime when an output dataset is being replaced by an analytic. In cases where scheduled big data analytics replace output data on every run, previously, users could experience periods where the layer was unavailable while it was being rewritten, which affected display in client applications. This has been addressed so that the previous version of the data continues to be available until the new output finishes being written.
An issue with geometry handling was fixed such that stream layers return Z information correctly and display successfully in 3D in ArcGIS Pro.
An issue with querying was fixed such that feature layers return correct results when using output statistics and returning distinct values.
An issue with output layer aggregation settings was fixed so that aggregation bins can be successfully defined with alternate spatial references.