Building SuburbIntel — An Australian Property Intelligence Platform

The Problem

Property investors in Australia face a fragmented information landscape. Government data is scattered across multiple sources — ABS Census, state valuer-general databases, infrastructure records, and development applications — none of which talk to each other. Investors end up spending hours cross-referencing spreadsheets just to evaluate a single suburb.

I wanted to build something that brings all of this together in one place: a platform where you type in a suburb name and instantly get a complete picture — demographics, growth trends, rental yields, infrastructure projects, and a clear investment score.

What I Built

SuburbIntel is a production SaaS platform built with Next.js 15 (App Router), React 18, TypeScript, Tailwind CSS, and PostgreSQL with Prisma ORM. It covers over 14,500 Australian suburbs with real, verified government data.

The platform includes 44+ pages featuring:

• Interactive heatmaps showing investment potential across regions
• Suburb comparison tools for side-by-side analysis
• Investment calculators for yield, growth, and affordability
• Real-time market analytics and trend visualizations
• A proprietary investment scoring engine (0-100) with weighted algorithms
• AI-powered suburb summaries and buyer persona generation
• An agent marketplace with featured placements and lead management

Data Engineering & ETL

The backbone of SuburbIntel is a custom ETL pipeline that ingests data from five government sources:

• ABS Census 2021 — 1.4GB+ across 119 data tables (demographics, income, housing, employment, transport)
• NSW Valuer-General — Property sales and valuations
• VIC Open Data — Victorian property market data
• Infrastructure databases — Planned and current projects
• Development applications — Council-approved building activity

I built custom CSV/XLSX parsers with fuzzy suburb-name matching and geographic validation to handle inconsistencies across data sources. Processing 15,000+ suburbs worth of census data alone required careful batching and error handling.

AI Integration

The platform integrates OpenAI GPT-4 for intelligent features:

• AI-powered suburb summaries that synthesize complex data into plain English
• Investment risk analysis and buyer persona generation
• A RAG (Retrieval Augmented Generation) system with vector embeddings stored in PostgreSQL for contextual property Q&A

Security was a priority — I implemented comprehensive AI safety measures including prompt injection detection (30+ attack patterns), input/output sanitization, and role manipulation blocking.

Cloud Infrastructure

SuburbIntel runs on Google Cloud Platform with an architecture built for reliability and scale:

• Google Cloud Run with auto-scaling (1-10 instances, 2 vCPU/2GB each)
• Multi-stage Docker builds with standalone Next.js output
• CI/CD via GitHub Actions: TypeScript checks → Playwright E2E tests → Docker build → Artifact Registry → Cloud Run deploy → smoke tests
• Secrets managed through GCP Secret Manager
• Database on Supabase with PgBouncer connection pooling

Security Engineering

Given the platform handles user data and financial analytics, security was built in from the start:

• Database-backed rate limiting with tiered limits per endpoint type, supporting horizontal scaling
• Input sanitization middleware covering XSS, SQL injection, CSV injection, and path traversal
• Webhook idempotency for Stripe event processing
• Comprehensive AI prompt injection detection and blocking

Testing & Quality

Quality assurance covers multiple layers:

• 13 Playwright E2E test suites covering cross-browser testing (Chromium, Firefox, WebKit), mobile viewports, API endpoints, data quality, performance/SEO, and visual regression
• Type safety across 30+ Prisma data models and 50+ API routes with Zod schema validation
• Automated smoke tests as part of the deployment pipeline

Agent Marketplace

Beyond analytics, I built a real estate agent marketplace with signup, verification, listings CRUD, lead management, and admin moderation. The system includes featured placements with CPC/impression billing via Stripe integration.

Key Numbers

Tech Stack

Try It Yourself

SuburbIntel is currently free and open for everyone to use — no account required. Sign-in, signup, and Stripe payments are disabled so you can explore the full platform without any barriers. Search any Australian suburb, compare neighbourhoods, and see the investment scoring in action.