Geofencing and geospatial aggregation scenarios with Azure Stream Analytics

With built-in geospatial functions, you can use Azure Stream Analytics to build applications for scenarios such as fleet management, ride sharing, connected cars, and asset tracking.

Geofencing

Azure Stream Analytics supports low latency real-time geofencing computations in the cloud and on the IoT Edge runtime.

Geofencing scenario

A manufacturing company needs to track assets on their buildings. They equipped every device with a GPS and want to receive notifications if a device leaves a certain area.

Reference data used in this example has the geofence information for the buildings and the devices that are allowed in each of the buildings. Remember that reference data could either be static or slow changing. Static reference data is used for this scenario. A stream of data continuously emits the device ID and its current position.

Define geofences in reference data

A geofence can be defined using a GeoJSON object. For jobs with compatibility version 1.2 and higher, geofences can also be defined using Well Known Text (WKT) as NVARCHAR(MAX). WKT is an Open Geospatial Consortium (OGC) standard that is used to represent spatial data in a textual format.

The built-in geospatial functions can use defined geofences to find out if an element is in or out of a specific geofence polygon.

The following table is an example of geofence reference data that could be stored in Azure blob storage or an Azure SQL table. Every site is represented by a geospatial polygon, and every device is associated with an allowed site ID.

Generate alerts with geofence

Devices can emit their ID and location every minute through a stream called DeviceStreamInput. The following table is a stream of input.

You can write a query that joins the device stream with the geofence reference data and generates an alert every time a device is outside of an allowed building.

SELECT DeviceStreamInput.DeviceID, SiteReferenceInput.SiteID, SiteReferenceInput.SiteName 
INTO Output
FROM DeviceStreamInput 
JOIN SiteReferenceInput
ON st_within(DeviceStreamInput.GeoPosition, SiteReferenceInput.Geofence) = 0
WHERE DeviceStreamInput.DeviceID = SiteReferenceInput.AllowedDeviceID

The following image represents the geofences. You can see where the devices are in accordance to the stream data input.

Device "C" is located inside building ID 2, which is not allowed according to the reference data. This device should be located inside building ID 3. Running this job will generate an alert for this specific violation.

Site with multiple allowed devices

If a site allows multiple devices, an array of device IDs can be defined in AllowedDeviceID and a User-Defined Function can be used on the WHERE clause to verify if the stream device ID matches any device ID in that list. For more information, view the JavaScript UDF tutorial for cloud jobs and the C# UDF tutorial for edge jobs.

Geospatial aggregation

Azure Stream Analytics supports low latency real-time geospatial aggregation in the cloud and on the IoT Edge runtime.

Geospatial aggregation scenario

A cab company wants to build a real-time application to guide their cab drivers looking for ride towards the areas of the cities currently experiencing higher demand.

The company stores logical regions of the city as reference data. Each region is defined by a RegionID, RegionName, and Geofence.

Define the geofences

The following table is an example of geofence reference data that could be stored in Azure blob storage or an Azure SQL table. Every region is represented by a geospatial polygon, which is used to correlate with the requests coming from streaming data.

These polygons are for reference only and do not represent actual city logical or physical separations.

Aggregate data over a window of time

The following table contains streaming data of "rides."

The following query joins the device stream with the geofence reference data and calculates the number of requests per region on a time window of 15 minutes every minute.

SELECT count(*) as NumberOfRequests, RegionsRefDataInput.RegionName 
FROM UserRequestStreamDataInput
JOIN RegionsRefDataInput 
ON st_within(UserRequestStreamDataInput.FromLocation, RegionsRefDataInput.Geofence) = 1
GROUP BY RegionsRefDataInput.RegionName, hoppingwindow(minute, 15, 1)

This query outputs a count of requests every minute for the last 15 minutes by each region within the city. This information can be displayed easily by Power BI dashboard, or can be broadcasted to all drivers as SMS text messages through integration with services like Azure functions.

The image below illustrates the output of the query to Power BI dashboard.

Next steps

• Introduction to Stream Analytics geospatial functions
• GeoSpatial Functions (Azure Stream Analytics)