Amazon Product Price Prediction

Leveraged BeautifulSoup, Selenium, and Keepa API to predict price changes using LSTM-RNN and sentiment analysis

PythonTensorFlowScikit-learnNumPyPandasBeautifulSoupSeleniumKeepa API

Built a comprehensive machine learning system to predict Amazon product price changes by combining historical pricing data with sentiment analysis from customer reviews. The project involved scraping over 60 Amazon products and analyzing 4000+ reviews to identify patterns that precede price fluctuations.

Technical Highlights:

  • Implemented VADER algorithm for sentiment analysis on Amazon reviews to detect sarcastic patterns
  • Built an LSTM-RNN model with early stopping to predict price changes based on review sentiment
  • Achieved significant accuracy in predicting price movements prior to actual changes
  • Presented research findings at the Undergraduate Research Symposium to 200+ researchers

The model successfully identified correlation between review sentiment patterns and subsequent price adjustments, providing valuable insights for both consumers and retailers.