Creating a Real Time Dashboard (RTD) using Copilot
Why create a Real-Time Dashboard with Copilot
In typical Fabric / RTI implementation a user will have access to many sources of data that can help the user to make decisions about the business. Sometimes it can be overwhelming looking at a long list of streams and tables, not knowing how to get an overview and quick insight from the table’s contents. For that scenario we enabled creating Real-Time Dashboards based on tables in KQL databases.
We have two goals in using Copilot for this task:
- Make the initial creation of a dashboard fully automatic without requiring any technical knowledge.
- Educate the user about techniques that can be used when building dashboards like parameters and base queries.
Creating the dashboard
The first entry point for creating an RTD is the real time hub.
From a table you can click on the ellipsis and choose Create Real-Time dashboard
You’ll see a dialog and if you are ok with AI generated content you can select ‘Get started’.
You can stop the process while the dashboard is generated. The dashboard will be created as a new artifact in the WS in which the database that contains the selected table resides. If you don’t have the access rights to add anew artifact, the dashboard will be created in your personal Workspace. After a few seconds you’ll find yourself in a dashboard with two pages.
Copilot insights page
This page contains five tiles, the upper tile contains some links and necessary compliance warnings that this is AI generated content. The bottom 4 tiles are based on 4 KQL queries generated by the Copilot based on the selected table metadata and some general instructions for how best create KQL.
The instruction for the AI component is to create queries that will return interesting insights about the data. After we receive the queries we replace the name of the table with the base query – _Base. Base queries will be covered below when the profile page is explained.
By connecting all queries to the base query, it means that the two parameters of time range and by date are used in all tiles and you can limit the time range by any datetime column in the table including ingestion_time().
Profile page
The profile page is identical for all tables and includes 6 tiles, 4 base tables and 3 parameters. You can select any DateTime column to filter the displayed data. The selection will be used in both pages.
Locate the list of columns and select a column to see more details about it by using Cross-filter.
Summary
Real-Time Dashboards are a powerful tool for analyzing streaming data and timeseries data. To shorten the learning curve and provide quick insight and profiling the data, we use Copilot to create dashboards over a single table in a KQL database.
We have plans for enhancing this feature to use multiple tables and enhance existing dashboards with suggestions from Copilot.
Please send us feedback and ideas for improvements at Microsoft Fabric Ideas.
What’s next?
There will be many more exciting developments as we continue to innovate and expand the capabilities of Real-Time Intelligence. Learn more about all the features and follow a step-by-step tutorial. Join the conversation and vote for your favorite features.
Over the next month we’ll be releasing a series of blog posts that dive into all the capabilities further. Stay tuned for more!
