Insiders Hub

Sensible Shopper

Debdyut Hajra

Debdyut Hajra

Unknown

5 0
  • 0 Collaborators

The increasing ease of online shopping and the aggressive marketing policies of e-commerce giants are constantly driving more and more people to shopaholism. People frequently face the dilemma of whether or not to buy a product. Product pages and regular discounts are carefully designed to incline customers into buying a product as opposed to make an informed decision based on the need and want of the customers. Hence, an application is required which can help people to avoid the traps set by the companies and help them to make a more sensible decision before buying a product. ...learn more

Project status: Concept

Mobile, Artificial Intelligence

Groups
Student Developers for AI

Intel Technologies
AI DevCloud / Xeon, Intel Opt ML/DL Framework

Overview / Usage

Problem

Besides emotional reasons, there are many factors which lead shoppers to shopaholism. Some of them are as follows:

  1. Strategically designed product pages incline customers' decision to buy a product.
  2. Regular discounts attract customers to buy a product.
  3. Lack of channel to seek advice from peers and friends on the purchase of a product.
  4. Lack of proper methodology to verify the need and want of a customer with respect to a product.

Proposed System

The proposed application will help the customers in the following ways:

  1. Automatically analyze the sentiments from customer reviews across the internet.
  2. Ask yes/no questions to the user to verify his needs and wants with respect to a product.
  3. Send a survey to friends in contact to seek their advice and do post analysis on the retrieved data.

The above 3 metrics can help the users to check their shopping spree and help them to make an informed decision.

Methodology / Approach

It will be a web and Android application. There will a few basic steps in the application:

  1. The user can specify the product he/she is intended to buy, either by uploading a picture or proving a product link or just by typing a few keywords.
  2. The app will identify the product from the user input, and search for the corresponding product features and reviews throughout the internet. It will make a sentiment analysis on the reviews and ratings and display the statistics to the user.
  3. If the user is content with the acquired statistics, then the app will ask questions to verify that the customer actually needs the product and that if he/she should buy it at that time.
  4. It will also provide the option to seek advice from people in the user's contact list in the form of a short survey. It will process the surveys automatically and suggest whether or not to buy the product.
  5. Based on the above 3 metrics the user can then make an informed decision of purchasing the product.
  6. The result will be added to the database of the application to improve the efficiency of the system.

Technologies Used

  • The backend of the application will be developed completely using Intel distribution of python.
  • If a user uploads a product image, deep learning classifiers built using Intel optimised TensorFlow will be used to identify the product from the image.
  • Python web scraping libraries like BeautifulSoup and scrapy will be used along with urllib2 to extract user reviews and ratings from product pages.
  • Python NLTK toolkit will be used for sentiment analysis on the customer and peer reviews.
  • Django web framework will be used to develop the web API and the website.
  • Intel DevCloud will be used to develop and test the application.
  • Android Studio will be used to develop the Android application.
Comments (0)