Sentiment Analysis Accuracy

[KALAKUNTLAS]

Hi,

I am using Power BI Premium per user. I have data with the text of 2617 rows. I ran SENTIMENT ANALYSIS (text analytics in transform data ribbon) and got results. I didnt use any cognitive services. These results are not accurate. I think approx 30-50% error. Below is an example for one text.

I am wondering if anyone has experienced the same and found a way to reduce the errors.

[DataChant]

Queries with AI Insights will stop refreshing on September, 15, 2025. If you want to keep the AI Insights logic inside Power Query, read my blog post here, to learn your options and how to solve it.

The article demonstrates how to implement sentiment analysis using the new Azure AI Language service, and includes a reusable M function and example solution PBIX file for further learning.

[v-luwang-msft · Accepted solution]

Hi @KALAKUNTLAS ,

 Microsoft Cognitive Services API allows only 1000 text messages in a single API request call. If you have more than 1000 rows stay tuned to the blog post here with a detailed walkthrough that will also show you how to perform multiple API calls. 

And you can also refer the following link:

https://community.powerbi.com/t5/Community-Blog/Sentiment-Analysis-in-Power-BI/ba-p/55898 

Did I answer your question? Mark my post as a solution!

Best Regards

Lucien