AI Feedback Classifier using AWS Comprehend & Boto3

Project Overview

The AI Feedback Classifier is a sentiment analysis tool built using AWS Comprehend and Python. It processes customer reviews and classifies each one as Positive, Negative, Neutral, or Mixed. This project was part of my hands-on learning journey while preparing for the AWS Certified AI Practitioner exam. My goal was to move beyond passive study methods and start building real-world applications using AWS AI services.

This classifier demonstrates how Natural Language Processing (NLP) can be used to quickly extract customer sentiment at scale, making it useful for product teams, customer service, and business analysts.

Key Features

  • Parses and processes a CSV file of customer reviews

  • Uses AWS Comprehend’s detect_sentiment API

  • Classifies sentiment into four categories: Positive, Negative, Neutral, or Mixed

  • Displays sentiment alongside original customer text

  • Built entirely using Python, Boto3, and AWS CLI tools

Tech Stack

  • Programming Language: Python

  • Cloud Provider: AWS

  • AWS Services Used:

    • Amazon Comprehend

    • IAM (for permissions)

    • S3 (optional, for storing datasets)

    • CloudWatch (optional, for monitoring)

  • Libraries & Tools: Boto3, CSV module, AWS CLI, GitHub

Why I Built This

While studying for the AWS Certified AI Practitioner exam, I realized that I was falling into the trap of over-preparing with notes and flashcards but not actually building anything. I created this project to help solidify my understanding of how cloud-based AI services work in practice. It also served as a great introduction to working with real customer data and understanding the sentiment behind it.

How It Works

  1. I downloaded a dataset of women’s clothing e-commerce reviews from Kaggle.

  2. A Python script reads the dataset and processes each review using AWS Comprehend.

  3. The detect_sentiment() function returns the sentiment classification for each review.

  4. Results are printed to the terminal alongside the review text.

  5. This process can be extended to save results to a CSV file or a dashboard.

Challenges & Learnings

Setting up AWS permissions and configuring the CLI for the first time was a bit tricky. I ran into authentication errors and had to troubleshoot my IAM roles and access keys. I also learned how to properly parse large datasets with Python and how to handle encoding issues when working with CSV files. More importantly, I gained a better understanding of how AWS Comprehend fits into the broader world of NLP tools and how businesses can use these services to automate feedback analysis.

GitHub Repository

Want to see how it works under the hood? Check out the full project here:
👉 AI Feedback Classifier GitHub Repo

person writing on white paperperson writing on white paper