- web scrapping Farfetch.com fashion products
- create an end-user app with Deep Learning and Data Engineering recursively updating database
Can my model identify the bag’s brand based on the product image?
To use cloud computing and big data pipelines to create an e-commerce/fashion products images classification(to identify its brand) based on the product images given.
The data of product information will be web scraping from Farfetch.com by using API to collect product descriptions and images, etc. Next, I'll create a database to store the data on MongoDB.
I plan to start with one specific category (as Bag Section) to build up my model. I’ll take the images of 5 major bags brands – YSL, Gucci, Prada, Hermes, LV for modeling. I'll apply keras deep learning to train the model, then make it to a predict function to take a new image input to identify the bag’s brand. If necessary, I'll use Google Cloud Platform and Google Colab for a more efficient way to deal with big data.
It will be an end-to-end project, I aim to set up the pipeline for the whole process and deploy it into a web application for end-users.
- data collection: API json
- database management: MongoDB
- cloud computing: Google Cloud Platform & Google Colab
- image classification: Deep Learning keras
- app deployment: streamlit
- pipeline setup to get the baseline model running