Inspiration
After a full work week, my kitchen is usually filled with big amount of unused food waste. Even more, I have a collection of electronics sitting in my apartment for a while with no place else to go. Recently, my mom discovered that our local farmers market accepts food scraps, which made me wonder: are there other places in the city to dispose of this special type of waste? Turns out, there are a couple of NYC Open Data resources that map out these exact locations for disposal and smart composting bins.
New Yorkers are always on the go. As a New Yorker, I know our time is valuable to us, so I wanted to use data to help fellow New Yorkers tidy their environment intentionally and reduce overall food/electronic waste in our households.
What it does
DumpSmart helps New Yorkers get rid of things that don’t belong in regular trash- food scraps and electronics- without having to dig through city websites to find where. A user opens the site and sees an interactive map of the five boroughs with every drop-off site plotted, food scrap composting locations, and electronics recycling sites for computers, TVs, portable devices, and home electronics. From there, they can filter by borough, toggle between food scraps and electronics, or search directly by site name or address. Clicking a pin on the map or a card in the results list opens a detail view with the address, operating hours, seasonal availability, hosted organization, and a direct link to get walking or driving directions, so someone can go from “I have old batteries and produce scraps sitting around” to a specific nearby drop-off point in a few clicks.
How I built it
I used AI coding tools throughout the ideation, design, and development process. I used FigmaMake early on to generate a rough mockup of the layout and interaction patterns, which I then blended with my own design direction. Claude Code was my main development tool; I worked through iterative prompts to build out the map, filtering, the detail card view, and UI refinements, going file by file until the structure matched the design. GitHub Copilot helped with smaller inline suggestions and debugging while I was in the editor. On the data side, the app pulls live from two NYC Open Data datasets (food scrap and electronics drop-off locations) through their public API, so there's no backend or database, just a static site fetching real city data directly in the browser. Here's the datasets I used:
Challenges I ran into
My main challenge was not overcomplicating the project. Although I would like to expand the use case for this application, I wanted to make sure I was building a simple and effective application. Also, I struggled to ensure that my AI prompts were effective in what I requested from a design perspective.
Accomplishments that I am proud of
A functional interactive map that displays specific location info that a user is looking for.
What I learned
It is really important to be intentional while building public-facing applications with AI. From the development side, providing proper context and specifying tasks is a crucial aspect of effective prompt engineering. From a user experience side, it’s important to ensure that AI-generated or ‘vibe-coded’ applications are actually usable, useful, and accessible. This means making sure to practice a “human-in-the-loop” approach to building applications.
What's next for DumpSmart
An AI-native flow that users can directly prompt into, like “I want to get rid of an old server this Thursday at 3 pm, what’s the closest location I can do that?” This would let the app parse natural language for item type, timing, and location, then cross-reference open hours and accepted materials automatically, rather than requiring the user to manually toggle filters and scan a list themselves.
Built With
- claude
- css
- figma
- html
- javascript
- leaflet.js
- nyc-open-data
Log in or sign up for Devpost to join the conversation.