Amazon Laptop Market Analysis

This project automates the process of scraping, analyzing, and summarizing customer reviews of the top-selling laptops on Amazon.nl, with a focus on extracting product-specific insights using GPT-4o.

---

🚀 Pipeline Overview

1. Scrape Bestselling Laptops

• Source: Amazon.nl Bestsellers - Laptops
• Data collected: Title, URL, price, rating, and ASIN.
• Output: data/amazon_top50_laptops.csv

2. Collect Reviews

• For each ASIN, fetch customer reviews using Selenium and BeautifulSoup.
• Output: data/laptop_reviews/{ASIN}.json

3. Summarize Reviews with GPT-4o

• GPT-4o is used to extract product-specific praises and complaints, excluding unrelated content (e.g. shipping, seller).
• Input: JSON reviews per product.
• Output: results/laptop_review_summaries/{ASIN}.json

4. Merge Summaries with Metadata

• Combines review summaries with price, rating, and ASIN from original CSV.
• Adds brand and model parsed from title.
• Final output: results/final_laptop_dataset.csv

---

💡 Highlight: GPT-4o Summarization

We use GPT-4o to:

• Parse dozens of user reviews.
• Extract only meaningful, product-specific comments.
• Output structured, interpretable summaries for business insight.

This approach avoids noisy data from non-product-related content and makes the dataset highly usable for ML or analysis.

You can use it for downstream analysis, dashboards, or ML pipelines.