Sentiment Analysis in Power BI – A Step-by-Step Guide

Sentiment analysis is a powerful technique used to determine the emotional tone behind a body of text. By leveraging sentiment analysis in Power BI, businesses can extract insights from customer feedback, social media, surveys, and other text-based data sources. Power BI allows integration with AI models, Azure Cognitive Services, and Power Query transformations to perform sentiment analysis. This guide will provide a comprehensive step-by-step explanation of implementing sentiment analysis in Power BI.

Step 1: Understanding Sentiment Analysis

Before diving into implementation, it’s important to understand sentiment analysis. It categorizes text into different sentiment types, such as:

Positive
Neutral
Negative

Sentiment analysis can be useful for analyzing:
✔ Customer reviews
✔ Social media comments
✔ Support tickets
✔ Surveys and feedback forms

Power BI supports sentiment analysis using various approaches, including:

Azure Cognitive Services (Text Analytics API)
Power Query & Custom Models
Python/R Scripts

Step 2: Data Preparation in Power BI

To perform sentiment analysis, you need a dataset containing textual data. Follow these steps:

1. Load Data into Power BI

Open Power BI Desktop.
Click on Home > Get Data.
Select the data source (Excel, CSV, SQL Server, SharePoint, etc.).
Click Load after previewing the data.

Ensure that your dataset contains a column with textual data, such as “Customer Feedback” or “Comments.”

2. Clean and Transform Data using Power Query

Click Transform Data to open Power Query Editor.
Remove unnecessary columns to keep only the relevant text field.
Use Remove Duplicates to eliminate redundant feedback.
Replace null values with placeholders like “No Comment.”
Click Close & Apply when done.

Step 3: Performing Sentiment Analysis using Azure Cognitive Services

Microsoft’s Azure Text Analytics API provides an easy way to perform sentiment analysis.

1. Set Up Azure Cognitive Services

Go to the Azure Portal (https://portal.azure.com).
Search for Cognitive Services and create a new instance.
Select Text Analytics as the service type.
Copy the API Key and Endpoint URL from the resource page.

2. Connect Power BI to Azure Text Analytics

Open Power Query Editor in Power BI.
Click Add Column > Invoke Custom Function.
Use the following function to call the API:

let
    apiKey = "Your_Azure_Text_Analytics_API_Key",
    endpoint = "Your_Azure_Endpoint",
    sentimentURL = endpoint & "/text/analytics/v3.1/sentiment",
    requestData = [documents = List.Transform(Source[Comments], each [id=Text.From(_), text=_, language="en"])],
    jsonBody = Json.FromValue(requestData),
    headers = [#"Ocp-Apim-Subscription-Key"=apiKey, #"Content-Type"="application/json"],
    response = Web.Contents(sentimentURL, [Headers=headers, Content=jsonBody]),
    parsedResponse = Json.Document(response)
in
    parsedResponse

Click OK and wait for Power BI to process the sentiment scores.
Expand the JSON result to extract Sentiment Score (ranges from 0 to 1, where closer to 0 is negative and closer to 1 is positive).
Click Close & Apply.

Step 4: Performing Sentiment Analysis using Python in Power BI

If you prefer to use Python, follow these steps:

1. Enable Python in Power BI

Go to File > Options & Settings > Options.
Under Python scripting, set up your Python environment.
Install the required Python libraries:

pip install pandas numpy textblob

2. Apply Python Sentiment Analysis in Power Query

Click Transform Data > Add Column > Run Python Script.
Enter the following Python code:

from textblob import TextBlob
import pandas as pd

# Convert Power BI data into a DataFrame
df = dataset 

# Function to calculate sentiment
def get_sentiment(text):
    return TextBlob(str(text)).sentiment.polarity

df["Sentiment Score"] = df["Comments"].apply(get_sentiment)

Click OK and Expand Columns to extract sentiment scores.
Click Close & Apply.

Step 5: Visualizing Sentiment Analysis in Power BI

Once you have the sentiment scores, you can create visuals to analyze the sentiment distribution.

1. Create a Sentiment Score Column

Add a new calculated column using DAX:

SentimentCategory = 
SWITCH(
    TRUE(),
    'Table'[Sentiment Score] < 0, "Negative",
    'Table'[Sentiment Score] = 0, "Neutral",
    'Table'[Sentiment Score] > 0, "Positive"
)

2. Create Visualizations

Sentiment Distribution: Use a Pie Chart to show positive, neutral, and negative feedback.
Sentiment Trends Over Time: Use a Line Chart with date as X-axis and sentiment score as Y-axis.
Word Cloud: Use a custom Word Cloud visual to highlight frequently used words in comments.

Step 6: Automating Data Refresh for Sentiment Analysis

Go to Power BI Service (https://app.powerbi.com).
Navigate to Datasets > Schedule Refresh.
Configure refresh settings based on your data source.

Step 7: Sharing and Embedding Sentiment Analysis Reports

Publish the report to Power BI Service.
Share with stakeholders by providing access permissions.
Embed the report into a Power BI Dashboard, PowerPoint, or Website.

Conclusion

Sentiment analysis in Power BI allows organizations to gain insights into customer opinions and emotions. By integrating Azure Cognitive Services, Python scripts, or Power Query, businesses can automate sentiment scoring and create interactive reports. Visualizing sentiment trends over time helps companies make data-driven decisions to improve customer satisfaction.

Bonus: Tags for Sentiment Analysis in Power BI

Power BI, Sentiment Analysis, Azure Cognitive Services, Text Analytics, AI in Power BI, Natural Language Processing, Customer Feedback Analysis, Power BI Python Integration, Text Mining, Data Visualization, Machine Learning, AI-Powered Reports, Business Intelligence, Power BI Dashboard, Automated Insights