NethAddress: AI Property Intelligence Engine

Go
DevOps
System Architecture
Data Engineering
Full Stack

A production-grade property analysis toolkit aggregating 35+ Dutch data sources, built from scratch in Go and hosted on self-managed infrastructure.

NethAddress Hero Image

Live Project | GitHub Repository

Project Overview

NethAddress is a comprehensive property intelligence toolkit designed specifically for the Netherlands real estate market, aggregating data from over 35 sources including cadastral, demographic, and environmental records. Built entirely from scratch to solve the fragmentation of Dutch property data, it serves ~100 daily active users with a 2% conversion rate, offering actionable insights and scores for developers and investors.

Key Features

  • Comprehensive Data Aggregation: Unifies data from PDOK, Kadaster, CBS, and environmental APIs into a single source of truth.
  • Intelligent Scoring Engine: Calculates ESG, Profit, and Opportunity scores (0-100) using complex weighted logic.
  • Production-Grade Reliability: Robust error handling and graceful degradation ensures partial profiles are returned even if some downstream APIs fail.

Challenges & Architecture

The system is built on a Modular Monolith architecture using Go (v1.23+) for high performance and strong typing, essential for handling complex data structures. To manage the latency of querying 35+ external APIs, I implemented a Redis caching layer with specific TTLs. The infrastructure is self-hosted on a Hetzner VPS using Coolify and Docker, with a custom CI/CD pipeline that injects build hashes for precise version tracking. This project bridged the gap between coding and engineering, teaching me deep lessons in DevOps and system design. For a deeper dive into the aggregator pattern and scoring logic, check out the GitHub Repository.