Housing Price Map

Our Housing App!
System Design

Inspiration

We wanted to tackle the CoStar Group Real Estate Data challenge, because we felt that through studying housing market data, we could potentially build a powerfully socially impactful app at Bitcamp.

What it does

Our app's CockroachDB instance holds around 20000 instances of housing data from the state of California. We've built a web interface for our users to query the average cost of a house within an specific location. The app intuitively visualizes the user's query parameters in real time through the use of map overlays, and heat maps.

How we built it

We wanted to leverage the most exciting technologies being provided to us at BitCamp, so naturally we gravitated towards a Next.js and CockroachDB stack hosted on Microsoft Azure. Much of our time at BitCamp was spent on making our CI/CD process as smooth as possible. Once we had built our Azure DevOps pipeline, we focused on implementing the core functionality of our app in an pseudo-Agile environment. Near the end of Bitcamp, as we become more satisfied with our app's functionality, we began to polish the UI/UX of our web app.

Challenges we ran into

We had much to learn about DevOps with Azure. The majority of our time was spent reading documentation to build our CI/CD pipeline. We also had trouble with the specifics of implementing our REST API. And finally, we had to all brush up on our PostgreSQL to create the appropriate queries for our CockroachDB instance. However, I think our greatest challenges were also the most memorable aspects of Bitcamp. Especially after we'd solved them,

Accomplishments that we're proud of

We successfully delivered our MVP without too many hitches, and with much time to spare.

What we learned

We took a deep dive into Microsoft Cloud, and came out much more well-versed in DevOps. We also learned the power of Next.js as a Front-end framework, especially in its ability to facilitate API calls and routing.

What's next for Housing Price Map

The Housing Price Map project has much potential. Chiefly, with better datasets, more overlays can be generated. Additionally, we really wanted to implement a machine learning algorithm into our project (XGBoost), but a lack of labeled data had stopped us in our tracks.

Much improvement can be made in our app's UX. Although we overall are happy with our delivered product's features, we feel that our app would benefit from a long term development cycle.