Brick City Grove

Inspiration

Newark is full of local businesses, events, jobs, and community resources, but finding them means digging through a dozen different sites, social pages, and flyers. Because it's hard to see what's nearby, a lot of residents and students end up spending their money outside the city. We wanted to make it easy to discover Newark from inside Newark and keep that money circulating in the community.

What it does

Brick City Grove is a community discovery platform for Newark residents, students, and visitors.

Interactive map: local businesses, events, and resources pinned across the city
Neighborhood heatmap: shows where activity is happening in real time
Forest Guide: a Gemini-powered AI chatbot you can ask things like "What's happening near me this weekend?"
Business dashboard: local businesses can post events and track their visits and economic impact
Accounts: sign in with Google or email, with separate roles for residents and businesses (business accounts are approved by an admin)
Languages: To accommodate the diverse languages the Newark Community speaks, the website is available in English, Spanish, and Portuguese.

How we built it

Next.js + React (on Node.js) for the frontend and API routes
Gemini API powers the Forest Guide chatbot
Tiger Data (Postgres) stores users, businesses, events, and visits. We used a hypertable for visit data and a continuous aggregate to roll up visits and spending per place per hour, which feeds the heatmap and the dashboard
Azure Maps for the interactive map and heatmap layer
Auth.js for Google and email/password sign-in
Azure App Service hosts the app, deployed automatically with GitHub Actions
GoDaddy for our domain
Tailwind CSS for styling

Challenges we ran into

Deploying to Azure: our first deploys kept failing. We fixed it by switching to a Next.js standalone build and shipping it to App Service as a zip.
Authentication: getting Google sign-in, email/password accounts, user roles, and admin approval to all work together on the newest versions of Next.js and Auth.js took a lot of trial and error.
Time-series data: we had to learn how to model visits as a hypertable and query aggregates fast enough for a live heatmap.
Scope in 24 hours: we planned more integrations than we could finish, so we focused on getting the core experience working well.

Accomplishments that we're proud of

A working, deployed app on a real domain
A live heatmap backed by real time-series data, not hardcoded values
An AI guide that makes exploring the city feel conversational

What we learned

How to use a time-series database (hypertables and continuous aggregates) for real-time analytics
How to plug an LLM into a real product experience
How to set up auth, roles, and a CI/CD pipeline under hackathon time pressure
How to work as a team and divide a full-stack project in 24 hours