-
-
Vehicles (Sorted by transfer fee from San Francisco)
-
Scanning Driver's License (Computer Vision and Machine Learning)
-
Car Details View Controller
Inspiration
Have you ever bought a car? If so, you know how frustrating it is to go back and forth to your dealership only to verify a handful of documents (such as Driver's License) that we think can obviously be done online (and yes, we've been in such situation, and we hate it!). Thus, we reimagined a system to solve that problem, and that's how the idea of VehicleBuy is born.
Click here to watch the pitch deck
What it does
VehicleBuy reimagines the way we purchase used cars by providing insightful car listings from CarMax's colossal vehicle inventory, intuitive data querying, and simplifying the document verification process using Machine Learning and Computer Vision.
How we built it
We did Web Scraping from carmax.com to collect cars and dealership data and fetched it to our Datastax Astra database, which we connected to our iOS app. By accessing both the live user location and the dealership coordinates, we used MapKit to determine the shortest possible routing and measure the distance to calculate the transfer fee (we use this to visually narrow down car transfer fees as well!).
To optimize camera performance, our iOS mobile app is written in Swift to create a smooth native experience for our users. By utilizing Computer Vision, users can scan documents with the device's camera and provide real-time dimensional feedback guiding lines using VisionKit to help users scan properly. Using the scanned image, we used Google Firebase ML Kit to recognize texts from the given document image, which we used for the verification process and to reject unclear images.
Accomplishments that we're proud of
Despite the difficulty of not being able to meet each other in person, we were able to coordinate and deploy a polished and useful application. We were able to utilize both Computer Vision and Machine Learning to provide real-time document dimensional feedback and text recognition.
Challenges we faced
Since CarMax has no publicly available APIs, we spent significant time in Web Scraping to retrieve the car listing items and cleaning the data to suit our app's needs.
What we learned
We should've planned regular checkpoint meetings to keep ourselves coordinated. A significant amount of time was wasted in waiting for another person to respond. Every mistake (including bug fixing and miscommunication) cost us a lot more time compared to if we were in an in-person hackathon.
What's next for VehicleBuy
We are planning to expand database with other car dealers. To expand our reach, we are going to extend support for Android and Desktop (Website). We will also try to utilize the CV and ML for vouchers, coupons, etc.
Log in or sign up for Devpost to join the conversation.