Public Amenity Locator

Overview

The Public Amenity Locator is a community-driven web app for identifying and mapping public amenities using AI-powered image recognition. Users upload photos of amenities, and the app verifies the type using a Hugging Face image classification model. The goal is to build a crowdsourced, self-verifying map of useful public infrastructure that’s often missing from traditional mapping services.

Features

  • AI-Powered Verification: Upload a photo and the app confirms the amenity type using Hugging Face’s image classification API.
  • 8 Amenity Types: Supports rubbish bins, dog waste stations, benches, postboxes, drinking fountains, bicycle racks, ATMs, and EV charging stations.
  • Flexible Upload: Drag-and-drop, file picker, or paste an image URL.
  • Geolocation: Captures GPS coordinates for mapping verified amenities.
  • Gamification: Points system, leaderboard, and achievements to encourage contributions.
  • Responsive Design: Works on both desktop and mobile.

Tech Stack

  • Backend: Node.js + Express
  • Frontend: Plain HTML/CSS/JS (static files)
  • AI: Hugging Face Inference API (google/vit-base-patch16-224)
  • Upload Handling: Multer

Challenges

The main challenge was ensuring data integrity in a crowdsourced environment. Integrating computer vision as a gatekeeper helped filter out incorrect submissions and maintain a high-quality dataset. Handling image uploads, geolocation, and real-time verification required careful coordination between the client, server, and AI API.

Lessons Learned

This project highlighted the value of AI for data validation in crowdsourced apps. By automating verification, the app can scale its dataset while maintaining trust and accuracy—making it a promising approach for hyper-local mapping projects.