Shopping is always a mixed feeling - the joy of shopping and the pain of choosing. Reading comments is helpful, but it's hard to read through all, especially when there are thousands of them. We don't think that we still should be bothered by these today. With Natural Language Processing and Machine Learning, shopping is pure joy.
What it does
Our project focuses on makeups (as a start). It used NLP and did Sentiment Analysis so that users can easily see how people FEEL about the product without reading the comments. Users can search a certain makeup product on our webpage and find out what the Sentiment Score the product gets from all customers!
How we built it
We chose Sephora as a source website for all makeup products and their comments. We used Python to scrape all product data, including product information and all comments. Then we used the Google Cloud Natural Language API to do Sentiment Analysis on the scrapped data and formulated a score representing how much people like the product. Lastly, we created a website using Wix, on which the product info and scores are displayed and users can search for products.
Challenges we ran into
Neither of us was familiar with front-end development so we had to learn it as we go. It's also hard to manipulate a huge dataset of comments.
Accomplishments that we're proud of
Finished it in time! We are proud that we did a full-stack development with the use of Google NL API in 24 hours.
What we learned
Front-end development; use of GCP and Google Natural Language API; UI design; full-stack structure design; teamwork ;))
What's next for P!CK
The usage could expand to all online shopping websites and all forms of products - Amazon, Booking, Expedia, Sak Fifth Avenue, etc.