NuHomes.co(m)

Homepage

Inspiration

It's hard to predict the future, but it's most within reach when looking at trends from the past. New housing construction is a fantastic indicator of growth within a city or area, and our team wanted to create an intuitive visualization of new housing construction across the United States.

What it does

Our application visualizes, relatively and absolutely, an area's amount of new construction since January of 2022. Our application also offers LLM chatbot assistance to quickly and easily learn more about the area a user is viewing.

How we built it

The app uses SvelteKit as a web framework, in this case serving as both our frontend and backend. This is supported by various tools, like Vite, allowing us to quickly create API routes and efficiently transfer our data, as well as plenty of flexibility in how we organized and constructed our components.

The app is hosted on Vercel, and we used Git for version control. We used MongoDB Atlas for our database, and used Perplexity as the LLM behind our chatbot.

Data was obtained through Zillow's research portal, available here: https://www.zillow.com/research/data/ and cleaned utilizing python/pandas.

Challenges we ran into

With the majority of our group being first-time hackathon participants, we weren't sure what to expect. We initially had some trouble deciding on a project, and throughout our project we had to be very careful to avoid scope creep.

As expected, JavaScript rarely gave us helpful error messages, so time spent tracking down exactly where we made a mistake was costly. JavaScript's asynchronous workflow also proved to be syntactically difficult at times.

Accomplishments that we're proud of

One of our group members completed a feature from data collection to cleaning, then creating an API call to retrieve it, and finally displaying and formatting on the front end. He had never even touched JavaScript before this event and being able to learn so much so quickly was extremely impressive. The two of us also greatly enjoyed coaching and pair programming with him throughout this experience and we found it highly rewarding.

What we learned

Our group learned so much during our time working together. We all come from diverse backgrounds and have different experience levels. We learned much more about many tools and technologies such as map libraries, SvelteKit, MongoDB, API Calls, Pandas, JavaScript, Perplexity, and overall full-stack development. But most importantly we improved as members of a development team.

What's next for NuHomes.co(m)

There are some more features we'd like to see, mainly an improvement in the quality and amount of data we show for each location on our site. We would also like to be able to do an asynchronous monthly pull of housing data so we can update the pool passively. We plan to add these features in the near future!