ChatGPT x Power BI | Use ChatGPT to Analyze Data in Power BI | TUTORIAL

Utilizing ChatGPT within Power BI

In this video, we will learn how to use ChatGPT within Power BI to analyze data. The goal is to see if ChatGPT can take a value from Power BI and return it.

Setting up the Data

• Open Power BI.

• Create a table with the category "Bike", "Drain", and "Computer".

• Add quantity values for each category.

Using Power Automate

• Use Power Automate to get results from ChatGPT.

• Create a new flow in Power Automate.

• Choose an instant cloud flow.

• Select the premium connector for ChatGPT.

• Get API key from personal view API key.

• Save the API key.

Making Requests

• Go back to documentation and select "Making Requests".

• Select "Completions" and do a post request.

• Use headers for content type and authorization.

• Enter the API key in the body section.

Using GPT-3 to Generate Text in Power BI

In this video, the speaker demonstrates how to use GPT-3 to generate text in Power BI. They explain the limitations of maximum tokens and show how to add rows to a data set.

Using GPT-3 for Text Generation

• The speaker explains that they used GPT-3 for text generation.

• They demonstrate how they pulled the completion out of a body.

• The request and completion can only be 4,096 so you have to make sure that you're not sending in too much data because it needs to return at least.

• After hitting save, the post will return but we need to get that data into Power BI.

Adding Rows to a Data Set

• The speaker explains that they couldn't return it into about a field in Power BI.

• Instead, they added rows to a data set.

• They recommend having a timestamp for historical data.

• For now, we're going to reply with just one value: "reply".

• We use JSON parse and get the scheme from documentation.

• We select "text" as our field and save.

Running Flow

• We call it "Reply" and allow it to run flow.

Setting up Power BI Datasets

In this section, the speaker explains how to set up Power BI datasets and enable direct query for PBI datasets.

Enabling Direct Query for PBI Datasets

• To connect to real-time data, direct query for PBI datasets must be enabled.

• Restart Power BI desktop after enabling direct query.

Connecting to Real-Time Data

• Connect to different data sources by clicking on the "Power BI datasets" button.

• Create a single field called "reply" and keep it as a table of sorts.

• Save the dataset once it has been created.

Triggering Flows in Power Automate

This section covers triggering flows in Power Automate and checking if they are working properly.

Triggering Flows

• Trigger flows by clicking on the trigger flow button.

• Check if flows are working properly by refreshing the data and ensuring that responses are being received.

Including Data from Grid in ChatGPT Responses

This section explains how to include data from grids in ChatGPT responses.

Including Power BI Data in Responses

• Include power bi data in responses by adding it into the string using JSON format.

• Use prequalified statements such as "Please tell me what's the highest value" or "Please tell me what's the lowest value".

Analyzing Data with ChatGPT

This section covers analyzing data with ChatGPT and making improvements.

Analyzing Data

• Analyze data by asking ChatGPT questions such as "Determine that wall costs $5 million".

• Narrow down data to certain aspects by using sliced data.

Making Improvements

• Make improvements by adding selection of prequalified statements.

Getting Additional Data and Insights

In this section, the speaker talks about how to get additional data and insights.

Changing HTTP Request

• The speaker changed the HTTP request to ask for identifying patterns and generating insights for specific data.

• This change provided a lot more information, giving analysis of the data itself.

• The speaker recommends trying out this method to gain insights from data.

Conclusion

• The speaker concludes by saying that this was Some Guy named Chris and wishes everyone a great day.