review page highlight
Amazon Shopping Assistant Chrome Extension
SAP cloud platform: SAP psql, rabbitmq, django service.
To install and give it a try, simply go to chrome://extensions and load the latest Chrome Extension from your directory. It is published on the chrome extension web store
What it does
The extension is designed to digest thousands of amazon.com reviews for shoppers:
- match review sentence to certain product attributes
- generate sentiment score for review sentences
- the experimental support is currently limited on laptop catagory
Feature demo on live amazon laptop product and review page:
Scroll down to review matched relevant review sentences on certain product attributes:
Click review sentence to visit the entire review in details:
Refering to the image above
How we built it
The frontend is based on React.js along with Redux for state management. The chrome extension API is used for browser support.
The backend is based on Django server for review crawler, leverage by Celery/Rabbitmq to parallel the execution of crawling and matching mechanism. The database system is based on PSQL.
Challenges we ran into
When we are deploying our services to SAP cloud platforms, we first thought we can use full unix operating systems and command line tools to run them. However, we later found we can only use manifest files to deploy services. So we read the documentation and learned this new way of deployment during our hacking, and successfully built up all our services eventually.
Accomplishments that we're proud of
- speed up the crawler to control the return of first response under 60s
- scaled up the crawled for some product of 1000+ reviews
- learned a new way of software deployment that complies to the industry standard
What's next for Amazon shopping assistant
- [ ] setup user profile and feedback mechanism
- [ ] support more product categories
- [ ] .......