Objective

To compute summary statistics for each product based on sentiment predictions, rating data, and recommendation likelihood. This processed summary is then used to identify:

• Top 3 best-selling products by category.
• Bottom 3 least-selling products by category.

Data Source

• Input File: ../data/predicted_sentiment.csv
• Output Files:
  • ../data/summary.csv
  • ../data/output_best_selling.csv
  • ../data/output_least_selling.csv

Step-by-Step Process

1. Load Data

• Read the predicted sentiment dataset using pandas.

2. Create Summary Statistics per Product

• Grouped by product and categories.
• Aggregated the following:
  • predicted_sentiment: mean
  • rating: mean and count
  • doRecommend: mean
  • feedback: first sample for reference
• Renamed columns for clarity:

3. Clean Product and Category Columns

• Title-cased all product names.