AI in BI?! Yes, please.

Tips and Tricks of how you can level up your dashboards with AI

1. Using Column from Examples : Let's say you want to remove "out of 5" from the highlighted column, sure you can use split columns/extract first character but more simple than that? Use Column from examples and type "3" like in the first column and rest of the rows will follow. Just one click and you are done!

   
   

   This column detects pattern in your data and AI automatically understood the context of data. It also creates a PowerQuery itself.

   
   

   

   Using Column from Examples in PowerBI

2. Text Analytics: You can use Text Analytics to detect language, extract Key phrases and score sentiment

• Detect Language: Detects language of the text

• Extract Key Phrases: Sift through important phrases in your data (like reviews)

• Score Sentiment: Measure positive and negative sentiments of words or phrases. Will then generate a score out of 100.

3. Vision: If your data has a table with a list of image URLs, will look through a pattern in each image and generate tags based on that. Could be helpful in cataloguing and categorizing images without having to manually look through them.

4. Another AI tool that you can use in your visualizations is this:

   

   You can click on a certain data point and select Analyze ->explain the increase/decrease

This is how it would like:

If you like a certain visual you can click on + button and add it in your dashboard too

5. Smart narrative: This will analyze your visual and summarize what the changes were at a high level and highlight any anamolies in your data using NLP dynamically. You can also add this as a tooltip(but this doesn't get updated based on your selections in the dashboard). This changes as per your selection in the dashboard.

   You can customize smart narratives as well. You can add "dynamic values" which you want to include in smart narratives. You can also update what Smart Narrative has generated in your own language by referring to the auto calculated values (which smart narrative has created on the backend)

   
   

This is how it would look like if you click on smart narrative on the chart:

A quick text summary of visuals and reports is available with Smart Narrative

Smart Narrative Documentation

6. Q&A feature: You can enable Q&A feature and add "synonyms" to tag and categorize questions better

   

   
   

7. Key Influencers: Create a summary graph and highlights which analyzes your data and add the key influencers based on pattern

   
   

   

   You can add in the fields you want to Analyze and Explain by

8. Decomposition Tree: Helps you visualize data in a tree format, drilling down into categories (in this example) giving you breakdown

   You can select the category which you want to start with or you can select "High Value" which is a category created by the AI model that determines highest value for the metric. If it has a "bulb" infront of the category name, it means it is an AI Capability

   

9. Anomaly Detection: Under Analytics section in "Visualizations" you should be able to see "Find Anomaly" (if you have Copilot premium) and can update Sensitivity levels for the pre-trained models to highlight the points with possible explainations. Only works with Time-Series data.

   
   

10. Auto clustering:

    

    You can cluster data points based on proximity

11. Premium Sandbox insight generation: Pretty cool feature where you can analyze your dashboard as a whole and generate quick insights.

    
    

Video Referenced