Text analysis that returns the information grouped by topic

[Cin]

Hello,

I have one table of a survey and I have two columns that are open responses. I would like to be able to do an analysis of the responses (Text)  in which they can be grouped by subjets or topics.

The thing is that I don’t know the subjets or topics previously, the Power Bi have to do it by itself.

Do you know there is any visualización or if I can do it in any other way?

Thank you  in advance,

[Cin]

Hi,

Thank you for your answer. I send you an example of what I mean.

I have a table with a question and open responses:

Open Question p. ej your opinion about the trip

the trip was long

It was too expensive

the food was expensive

The bus ride was very long

the best, the friends I made

I really enjoyed the new friends I met.

And I would like to obtain a new column with the subjet, that can group the answers of the previous table:

Open Question p. ej your opinion about the trip	Subjets
the trip was long	Long trip
It was too expensive	Expensive
the food was expensive	Expensive
The bus ride was very long	Long trip
the best, the friends I made	Friends
I really enjoyed the new friends I met.	Friends

I hope you can understand now what I'm tryng to look for.

Thank you!

[Anonymous · Accepted solution]

Hi, @Cin 

Based on your description, you want to analyze survey responses in Power BI by automatically grouping them into topics or themes. So far, Power BI does not have built-in functionality for automatically analyzing text data and categorizing it into topics. You can submit an idea for here and the product team will take note of such a need.

Related Link: New tab (powerbi.com)

However, this can be done by integrating Power BI with Azure Cognitive Services or by using a custom script in Python or R to perform text analysis before importing the results into Power BI for visualization.

Here are the steps on how to perform text analysis and group survey responses by topic:

1. Preprocess the survey data to clean up the text data in the Power Query editor by removing any unnecessary characters, spaces, converting all text to lowercase, and so on.

2. Use Azure Cognitive Services (specifically the Text Analytics API) to analyze the text and identify key phrases or topics. Alternatively, you can write Python or R scripts to perform topic modeling using libraries such as NLTK or Gensim in Python or the and package in R.

Azure Text Analytics API link: Azure AI Language | Microsoft Azure
3. Import the processed data into Power BI as a new dataset or merge it with an existing dataset and then create visualizations in Power BI.

Best Regards,
Yang
Community Support Team

If there is any post helps, then please consider Accept it as the solution to help the other members find it more quickly.
If I misunderstand your needs or you still have problems on it, please feel free to let us know. Thanks a lot!

How to get your questions answered quickly --  How to provide sample data in the Power BI Forum

[Anonymous]

Hi, @Cin 

Based on your description, not sure what you want to do for now, what do you mean by analyzing as you describe? You can share pbix files without sensitive data, datasheets, descriptions of your requirements in the form of images (if possible) and the logic to implement your requirements.

Best Regards,
Yang
Community Support Team

If there is any post helps, then please consider Accept it as the solution  to help the other members find it more quickly.
If I misunderstand your needs or you still have problems on it, please feel free to let us know. Thanks a lot!

How to get your questions answered quickly --  How to provide sample data in the Power BI Forum