Celestial Biscuit

Redefining online grocery shopping with Machine Learning and APIs

Comment

Problem Statement

"A basket that understands you".
Negozio AI aims to enhance the online e-commerce space by offering a better experience that is closer to the offline supermarkets we all love to fill our shopping carts at.

Inspiration

Two interesting facts about offline supermarkets that we noticed are

  1. The items that are similar and have a higher chance of being bought together are stacked on the shelves which are close to one another.
  2. There is staff available at all times to help you find what you're looking for and notify you about new offers.

APIs We Built On Our Own

We utilized these facts to create a smart basket that is powered by machine learning algorithms.

  1. Similar User API: The users are also recommended products based on the taste of similar users. This is done by training the self-generated dataset on the Nearest Neighbors algorithm. The model was deployed as an API using Python and Flask on Azure.
  2. Similar Product API: The users are also recommended the alternative items of the current item. This is done by grouping similar products into one category. A flask API deployed on Azure sends a random item as the alternate product from the category.
  3. New User API: The new users with no purchase history are recommended the highest rated product by other users. A flask API was deployed on Azure to send the list of top 10 products to the web app.
  4. Past Purchase API: To recommend products based on past orders history, the Apriori Algorithm of Association Rule Learning was used. The machine learning model generated on training the Market Basket Optimization dataset was deployed using Azure to generate the Flask API.

The Recipe Arena by Negozio AI

When the world crumbled to the pangs of the Covid-19 virus, homemade food triumphed over food that was ordered from restaurants. Numerous people started looking for more ingredients, new recipes and understood the satisfaction one gets when the food they cooked gets the applause of all. Negozio AI uses the MealDB APIfor a quick search of recipes through just an ingredient! And to top it all, you also get a direct connection to the video recipe if you are interested.

Live Connect using Agora.io API

To bring the touch and feel of offline supermarkets to life. We have added a feature where our customers can connect with our representatives on a video call live. This feature can be used for several purposes:

  1. Customers can get assistance with finding items online.
  2. Customers can get to know about new offers, discounted products, and best-selling items. It is true that these items can have a tag on them but word-of-mouth always has more weightage over a label that we added through code.
  3. Brands can send in their representatives to set up their dedicated booths to introduce customers to their products and boost up their sales.

Tech Stack

Team Celestial Biscuit has used HTML, CSS, and JavaScript to build our web app, Nodejs for backend purposes, ejs as the templating engine, Figma for designing and prototyping, flask and azure for building and deploying our custom made APIs, Python, Jupyter notebook, and PyCharm to write our machine learning algorithms, Agora.io to enable live-chat, Vercel to host the front-end of our live-chat arena, Cloudinary to host our images, Heroku to host our application and the back-end of our live-chat arena, Firebase for authentication, TheMealDB API for getting our recipes, Postman for testing, Github for project management, VSCode for coding our application and finally Premier Pro and After Effects for creating our demonstration.

How To Use Negozio AI as a tester

  1. If you want to test the APIs that are built for a user whose past purchase history is already available, login to the web app with the following credentials:
    Email: olduser@gmail.com Password: olduser
  2. If you want to test the APIs that are built for a user who is new to the web app, please login through the undermentioned credentials:
    Email: newuser@gmail.com Password: newuser
  3. Using Negozio AI on your local machine
    Step One: Clone the repository
    git clone https://github.com/Nishtha0801/Negozio-AI
    Step Two: Open your IDE and run the following commands one after the other
    npm install
    npm start

Business Prospects and Utilisation Scope

  1. Negozio AI can be easily used as a feature in the existing e-commerce applications to provide recommendations to the customers for different products on their website.
  2. It can also be used as a full-fledged e-commerce platform by businesses where brands can be invited to promote their products and offer coupons and discounts to the customers to increase engagement with the web application. This will enable e-commerce platforms to get sponsorship from brands.

Built With

Share this project:

Updates