Skip to content

Latest commit

 

History

History
96 lines (65 loc) · 5.27 KB

README.md

File metadata and controls

96 lines (65 loc) · 5.27 KB
Raw

VisionX-Socializing E-commerce

Follow the trend

❓ Problem Statement

Myntra aims to provide best in class services and products to all their customers. Social Commerce has many facets and one of the emerging areas is influencer led content communities and e-commerce. Hence, provide an innovative idea that can help narrow the gap between e-commerce and Influencers. How can Myntra build and scale the overall influencer ecosystem?

✈ Solution

A lot of consumers today start their shopping journey with seeking inspiration and before deciding on the products to purchase. This inspiration drives the customers to refer to online shopping platforms and look for the same or similar products. We propose a social media(twitter/Instagram) bot that can help the customers to get links and other related information for their favourite influencer’s wardrobe on Myntra, just by adding an commenting/tweeting with #myntra_discover

image

🔥 Mockups

✨ Flow

1. Twitter API

  • Twitter Stream Endpoint - statuses/filter, to stream tweets with #myntra_discover
  • We extract in_reply_to_status_id_str to which we reply our results
  • Twitter Reply Endpoint - statuses/update, to give replies to the captured tweet

2. Ximilar

  • In order to get relevant tags based on image we use Ximilar API.
  • Ximilar API use a fashion tag service to detect different tags related to the provided image URL.
  • We primarily use Sub-category (t-shirt, pants, jackets etc), Gender(Male, Female, Unisex)

3. Google Cloud Vision API

  • In order to get relevant colour based on image we use Google cloud vision api.
  • Vision API provides colour in rgba form.
  • This rgba is converted into hex code using rgb2hex package and then to colour name using color-namer package.

4. URL Formation

  • Myntra uses a special route for all categories along with gender.
  • Example: For “Men T-Shirt” category, Myntra uses “/men-shirt” route.
  • Myntra also uses URL query parameters to filter the results.
  • Example: For getting a “white color t-shirt for men. Myntra redirect to the “/men-shirt” route along with query parameters. “?f=Color%3A<ColorName><hex_value>” - Myntra uses query parameters to select sorting pattern as well
  • Example: For sorting the products based on popularity, Myntra adds “&sort=popularity”.
  • Hence, the final URL is https://www.myntra.com/men-tshirts?f=Color%3AWhite_f2f2f2&sort=popularity

5. Web Scraping

  • Scraping the products page of Myntra website and capturing the products in a JSON format.
  • Data could be found in 11th script tag of the page in window.__myx
  • Sample JSON is shown in the picture alongside.
  • The appropriate scrapped data will be posted as tweet/reply by the bot to the consumer/customer so that they can choose the product as per their preferred choice.

💡 Features

  • Auto Replies to the tweets with #myntra_discover with relevant product information present in the image of parent tweet
  • Top 5 Product recommendations according to tagged photo through notifications and replies
  • Product Recommendation with brand, ratings, sizes, discounts, MRP, price and a direct link to buy the product
  • Can be used by anyone eg. influencers, common public etc

⛷ Business Applicability

Because we are building on existing platforms such as Twitter, Myntra, and Instagram, the implementation of our solution is very smooth and simple. We are not a separate app or website that users must download or visit. We are a service that will be available on the most popular social media platforms, such as Twitter, Instagram and Facebook (future). When someone posts a photo of themselves, especially actors/creators/influencers, we as users want to buy a similar product, but it is difficult to find clothings related to that image. As you have seen our solution does that for you! Simply reply to the tweet with a hashtag, and our Bot will share the link and other product details as a reply or in your inbox messages in your inboxes. As a result, our solution is both time-saving for users and profitable for myntra.

🧠 Potential Impact & Future Scope

Our project targets to solve the problem statement in such a way that it creates ease for customers to shop resulting in more use of our service. This project would be the first small step towards the socialization of e-commerce companies. A new consumer survey indicates that roughly half of consumers respond to social media as a marketing and commerce channel. Keeping this data in mind our service can make a huge impact on socialization of e-commerce and on customer service as well. Future expansion of this service on other social media platforms like Facebook/Youtube/Pinterest will increase the overall traffic on Myntra’s website and the potential customers.

👣 Steps to run locally

  1. Clone the repo in your local machine and setup the node environment. Create .env file similar to .env.sample file with all the required fields.
  2. Do yarn in root directory to install all the dependencies.
  3. Type yarn dev in the terminal to start the development server.

🤖 Tech Stacks & dependencies

  • nodejs
  • web scraping
  • cloud vision api
  • ximilar
  • Twitter API
  • ExpressJS
  • Myntra