Inspiration
No money. GST hike. Broke.
What it does
Our project compares prices between supermarkets to find out which provides the best deal.
How we built it
We scraped product data from Cold Storage, Fairprice, Sheng Siong ~and Redmart~. We then tried to match equivalent products by cosine similarity and displayed it on a frontend.
Challenges we ran into
Scraping data is a very interesting task: different supermarkets have different ways to get data. Some provide websockets, some make us parse HTML each in their different formats and some expose their backend APIs. We also had to clean the data in a consistent format for the next step.
The biggest challenge we faced was trying to figure out which products were correlated. We solved this with
zichen's black magic functionsome nlp thing.
Accomplishments that we're proud of
- We completed it.
What we learned
- Scraping data might not always be perfect.
- Hard to prevent false positives.
What's next for price chope chope
- Improved matching, especially size.
- Automated daily CRON jobs for latest prices
Log in or sign up for Devpost to join the conversation.