A Real-Time Scraping Approach to a Comparison-Based E-Commerce Website

Abstract

In this project, we performed web scraping and collected data from E-commerce websites like Flipkart and Snapdeal with the help of well-known Python libraries like BeautifulSoup and Selenium. To design the website a Python web application framework known as Streamlit was used as it is very useful and easy to use. For the second part of the project, we used a dataset of 44,000 images which we downloaded from Kaggle to train a machine learning model in which the TensorFlow and ResNet50 model was used. Resnet50 is a well-known convolutional neural network (CNN). After training the model we included it in the Streamlit web app._x000D_ We made a user-friendly platform in which users can upload any image, and they will be recommended 5 similar products with the help of which they can have a wider option to choose and see what all they have in the market is similar to what they like. Also, there is an option to check the least price of products on Flipkart or Snapdeal which people can search and review. The product we give can be useful to people who think twice before buying any product because everyone wants options at a lesser price. It will save money and time for the users they can directly check the top 5 products with less price, and then they can check the variety of options they have in similar products.