Highlights

• A clone of the NYC parking app, built with a React client and Node.js/Express backend.
• Designed a new database schema and seeded it with publicly-available meter data to support the geo-spatial layer.
• Implemented k-means clustering & utilized data visualization to present the meter data.

About

Another fun personal project - I found myself looking at the NYC parking meter app constantly and wondering what it would take to build a working version. I realized that the data representing all the parking meters in New York City, including their geographic locations (lat/long coords) was publicly available as part of the OpenData project. Once I got my hands on that, it was just a matter of setting up the database and seeding it with meter data. From there I was able to build REST endpoints in a Node.js environment, utilizing TypeScript for end-to-end type-safety, and start sketching the Google Maps implementation in the client. The map view was the heaviest lift, but I also had to add models and handling around creating and extending (parking) sessions and managing the user's cars.

One of my favorite parts of this project was learning about the k-means clustering algorithm, which I had to implement in order to gather the different meters into groups, depending on the user's zoom settings and location.

This project is still in-progress.

Tags

Links