FlipAssistant.AI

Logo Banner

Inspiration

New eBay users looking to sell their old phones have a hard time finding out what they should sell their phone for. If they sell for too little, they lose out on potential profit, and if they sell for too much, then it takes forever to sell their phone. Also, phone recycling businesses give misleading quotes because they are incentivized to maximize their own profits. FlipAssistant.AI aims to fix this issue.

What it does

Flip Assistant.AI solves both ends of this problem by being the first tool to provide unbiased market prices and automate the searching process for listings. It benefits all buyers and sellers by automating their decision making processes.

How we built it

How it Works:

The program collects user input from site
Gathers listings from eBay in real-time
Trains clustering model (K-means) using collected data
Calculates cluster closest to user's input
Does weighted calculations to find market price of user's phone
If user wants to buy, it shows listings below market price
If user wants to sell, it shows the statistically most similar listings

Challenges we ran into

We developed an automatic web scraper, which fetches data from HTML code, converts it to a CSV file, and converts all the categorical values into numerical ones. This process is much more tedious than we expected.

Accomplishments that we're proud of

Implementing K-means from scratch allowed us to dig in the data and find the closest data points, which allowed us to offer the 'similar listings' and 'below market price' features.

What we learned

Web scraping and data collection is harder than it looks.

What's next for FlipAssistant.AI

The aim is to expand this process to support more tech than just iPhones, and also to speed up the data collection process, so this could be used for large (100,000) sized data sets.