CivicPulse

CivicPulse is an AI-assisted civic reporting platform designed to make reporting local infrastructure problems faster, easier, and more transparent.

The idea started with a simple problem: citizens often see issues such as potholes, garbage accumulation, broken streetlights, water leakage, and drainage problems, but reporting them can be fragmented and difficult to track.

What CivicPulse does

Citizens can:

• Report a local civic issue

• Add photos or short video evidence

• Share their location

• Describe the problem

• Receive AI-assisted issue classification and severity suggestions

• Check whether a similar unresolved issue already exists

• Support existing issues instead of creating unnecessary duplicates

• Track the issue and its updates

AI-assisted workflow

CivicPulse uses an AI layer to analyze a citizen's report and generate:

• Issue category

• Subcategory

• Short summary

• Possible severity

• Relevant keywords

• Confidence score

The system also checks existing unresolved reports to identify potential duplicate issues.

Evidence & transparency

Reports can include multiple pieces of image/video evidence. Submitted evidence is linked to the corresponding civic issue so that the report contains more useful context than a text-only complaint.

Technical implementation

Frontend:

React + TypeScript + Vite

Backend:

Node.js + TypeScript

Testing:

Vitest

The project includes validation, authentication, rate limiting, duplicate detection, AI-assisted classification, evidence handling, and automated tests.

Why I built it

Civic problems are often visible to everyone, but the reporting process can feel invisible to the citizen.

CivicPulse explores how AI and a simple citizen-first interface can turn an observation on the street into a structured, traceable civic report.

Current status

CivicPulse is currently an MVP focused on the citizen reporting experience and the underlying reporting workflow.

Future direction

The next stage is to make the AI more capable of understanding visual evidence, automatically extracting the likely issue from uploaded images, improving severity estimation using location and context, and making the reporting process require as little manual input from citizens as possible.