What is the best solution for live dashboard using powerBI?

Aurovind Sagar Epari 80 Reputation points
2023-09-18T10:10:22.3266667+00:00

For a IoT solution for a device which sends tons of sensor data from more 10-12 sensors the device has, what is the best solution to have,

  • A live dashboard showing last sensor values
  • device state
  • Live dashboard showing 1 hour of sensor values
  • Historical dashboard for 1 month
  • I also have some requirements for alerts.

I am planning to have a stream analytics with IoTHub as input and PowerBI as output. And finally a IoT dashboard webpage embedding PowerBI dashboard and other features that we need.

But I am thinking is, I can skip stream analytics by querying data from PowerBI (direct queries) even for live dashboard. In what other ways stream analytics can be helpful? Is it fine I ditch stream analytics, a good idea?

I know the question is bit broad in nature but any suggestions would be helpful before I jump into detailed implementation.

Azure IoT
Azure IoT
A category of Azure services for internet of things devices.
384 questions
Azure Stream Analytics
Azure Stream Analytics
An Azure real-time analytics service designed for mission-critical workloads.
333 questions
0 comments No comments
{count} votes

Accepted answer
  1. Sander van de Velde 29,286 Reputation points MVP
    2023-09-18T11:07:22.37+00:00

    Hello @Aurovind Sagar Epari ,

    there are many ways to build an IoT solution, the one you describe is a following a very common pattern. One of the big advantages of adding ASA to it is that you can easily integrate alerting because you have control over the live stream.

    If you put the (transformed) data in a database first, before you query it using PowerBI, you have more control over the data. For example, an alternative is using Azure Data Explorer (there is even a free version available to try it out).

    Put your dashboards on top of that database.

    Regarding the device state, you need to ingest non-telemetry messages from the IoT Hub and store them along with your ingested telemetry.

    Regarding the real-time dashboard, you could also think about using Managed Grafana.

    I recommend trying to keep some flexibility (eg. using the Stream Analytics job) so you can learn and change the logic without too much impact.

    Because you have 'tons of sensors data', try to calculate the usage/ cost impact. Keep an eye on the load of your services.

    You can always migrate towards custom logic in eg. Azure Functions if needed.


    If the response helped, do "Accept Answer". If it doesn't work, please let us know the progress. All community members with similar issues will benefit by doing so. Your contribution is highly appreciated.


0 additional answers

Sort by: Most helpful