Tutorial: Analyzing Customer Feedback

In this tutorial, we'll walk through the process of analyzing customer feedback data using Probably. We'll focus on uncovering key trends, sentiments, and actionable insights from a dataset of customer reviews.

Dataset Overview

For this tutorial, we'll use a dataset of customer reviews for a fictional e-commerce platform. Each row in our dataset represents a single customer review and includes the following columns:

• review_id: Unique identifier for each review
• product_category: Category of the product being reviewed
• rating: Customer rating (1-5 stars)
• review_text: The full text of the customer review
• review_date: Date the review was submitted
• customer_id: Unique identifier for each customer
• purchase_verified: Boolean indicating if the reviewer purchased the product

Step 1: Loading the Dataset

1. Download the customer feedback dataset from [link to dataset].
2. Open Probably and click on "Add Dataset" in the top right corner.
3. Upload the CSV file containing the customer review data.
4. Review the auto-detected column types and adjust if necessary:
   • Ensure review_id and customer_id are set as categorical
   • product_category should be categorical
   • rating should be numeric (discrete)
   • review_text should be text
   • review_date should be datetime
   • purchase_verified should be boolean
5. Click "Confirm" to load the dataset.

Step 2: Initial Data Exploration

1. Create a frequency plot of the rating column:

   • Set X-axis to rating
   • Observe the distribution of ratings

2. Create a box plot with product_category on the X-axis and rating on the Y-axis:

   • This will give you an overview of rating distributions across different product categories

Step 3: Analyzing Sentiment Across Product Categories

1. Create a stacked bar chart:

   • Set X-axis to product_category
   • Set Y-axis to rating
   • This will show the proportion of different ratings for each product category

2. Use the Z-axis feature to highlight significant differences:

   • Set Z-axis to rating
   • Observe which categories have significantly higher or lower ratings

Step 4: Exploring Review Text

1. Switch to Cluster View:

   • Set X-axis to review_text
   • Observe the clusters that form based on the content of the reviews

2. Color-code the clusters:

   • Set Z-axis to rating
   • Observe how sentiment correlates with the clustered topics

3. Summarize clusters:

   • Use Probably's summarization feature to get an overview of the main themes in each cluster

Step 5: Analyzing Trends Over Time

1. Create a line plot:

   • Set X-axis to review_date
   • Set Y-axis to rating
   • Group by product_category
   • This will show how ratings for different categories have changed over time

2. Use the sliding window feature to smooth out daily fluctuations and observe overall trends.

Step 6: Investigating the Impact of Verified Purchases

1. Create a comparative box plot:

   • Set X-axis to purchase_verified
   • Set Y-axis to rating
   • This will show if there's a significant difference in ratings between verified and non-verified purchases

2. Use the filtering feature to focus on specific product categories and observe if the impact of verified purchases varies by category.

Conclusion

Summarize the key insights discovered:

• Overall sentiment distribution
• Product categories with the highest and lowest satisfaction
• Main themes in positive and negative reviews
• Trends in customer satisfaction over time
• Impact of verified purchases on ratings

Discuss potential actions based on these insights:

• Areas for product improvement
• Customer service focus points
• Marketing strategy adjustments

By following this tutorial, you've learned how to use Probably to analyze complex customer feedback data and extract actionable insights. This process demonstrates how Probably's intuitive interface and powerful visualization capabilities can streamline your data analysis workflow and provide valuable insights quickly and easily.