AI in Power BI Desktop: Analyze Text and Image

The latest AI features in Power BI Desktop make it easier to gain insights from unstructured data, like customer comments and images. In this post, we explain what these capabilities are, how to use them, and what you can learn from your data.

For expert-led Power BI training, visit: Power BI Training at DataBear

Why Use AI in Power BI Desktop?

Many datasets include valuable information in text or images  such as customer reviews or property photos  which traditional charts and tables cannot easily analyze.

The AI features in Power BI Desktop help you:

• Analyze sentiment in customer feedback.
• Detect key phrases and topics in comments.
• Tag and analyze images for patterns.

These insights allow you to go beyond numeric data and understand what customers are really saying or seeing.

Analyzing Customer Comments

Step 1: Prepare Your Data

In Power Query, load a dataset that contains a text column with customer comments or reviews.

Step 2: Enable Text Analytics

Text Analytics is a preview feature. To enable it:

• Go to Options > Preview Features in Power BI Desktop.
• Turn on AI Insights.

Once enabled, you can use Text Analytics in Power Query.

Step 3: Apply Sentiment Analysis

In Power Query, select the text column and open Text Analytics. Choose Sentiment Analysis.

The tool assigns a sentiment score between 0 and 1 for each comment:

• 0 = very negative
• 1 = very positive

You can also extract key phrases and detect the language automatically.

Step 4: Use Sentiment in Reports

Once the sentiment scores are added as a new column, you can create visuals showing:

• Sentiment over time.
• Sentiment by region or category.
• Comparison of positive vs. negative reviews.

Analyzing Images

If your dataset includes image URLs or metadata, you can also apply Vision AI to tag images and identify patterns. For example:

• Detect common themes in property photos.
• Correlate image content with customer sentiment.

This can highlight areas where visuals and customer experience are misaligned.

Example: Seattle Rental Listings

In the demonstration, a dataset of Seattle property rentals included both numeric and text data.

• Text Analytics was used to score customer reviews and extract key phrases.
• Vision AI tagged images, revealing that gardens and outdoor areas were often linked to lower sentiment.

By combining these insights, you can understand both what customers say and what they see.

Key Notes

• Text Analytics and Vision AI require Power BI Premium to run.
• Azure Machine Learning models can also be integrated, without requiring Premium.
• The features work in both Power BI Desktop and Power BI Service (Dataflows).

Get Started

To use these features:

• Update to the November 2019 release or later.
• Enable preview features in Options.
• Ensure you have access to a Premium capacity if using Text Analytics or Vision AI.

For full documentation and updates, visit the official Power BI blog.

Summary

The AI features in Power BI Desktop help you unlock insights from unstructured data. By analyzing customer comments and images, you can better understand customer sentiment, identify key issues, and improve decision-making.

For further learning, visit:
Power BI Training at DataBear