CheckHowMuch.sg

What it is

CheckHowMuch.sg is a property intelligence site that helps Singapore homeowners understand what their home is worth, whether it is an HDB flat or a private condo. Every HDB block and every condo project gets its own page with transaction history, price trends, valuations, and lease-decay analysis.

The site covers all 26 HDB towns and private condos across 28 districts, with individual pages for every block, project, and flat type. Each page shows actual transaction data from data.gov.sg for HDB and URA for private property, including median prices, price-per-square-foot trends, and lease-decay impact over time.

The problem it solves

Most Singaporeans rely on property agents or generic portals for HDB resale data. The information is scattered, often outdated, and always filtered through someone trying to sell you something. I wanted to build a tool where homeowners could look up their specific block, see real transaction history, and understand what their flat is actually worth, without talking to anyone.

Why I built it

I wanted to prove that one person with AI could build something that would traditionally require a team of developers, a data engineer, and months of work. CheckHowMuch.sg was that proof.

The conventional approach would involve hiring a frontend developer, a backend engineer to build the data pipeline, and a data analyst to structure the information. That team would take months and cost tens of thousands. I built the entire thing with Claude Code in weeks, for zero development cost.

How it was built

The entire site was built using Claude Code and Astro. No agency, no WordPress, no templates. I wrote the data pipeline, the page generation logic, and the SEO structure through conversations with Claude Code.

Data pipeline

The pipeline pulls HDB resale data from data.gov.sg and private property transactions from URA, together covering 357,000+ transactions. A Python script processes and structures the data by town, district, block, project, and flat type. The output feeds into Astro’s static site generation, producing a unique page for every combination.

Page generation

Astro generates over 11,000 static pages at build time, one for every HDB block, condo project, town, and district. Each page is unique, with its own content, charts, and metadata. The site architecture is fully interlinked: town and district pages link to block and project pages, those link to flat type breakdowns, and every page links back up the hierarchy. This internal linking structure is critical for SEO.

SEO from day one

Every page has proper schema markup (WebPage + Dataset), unique meta descriptions, canonical URLs, and Open Graph tags. The sitemap is auto-generated. AI crawlers (GPTBot, ClaudeBot, PerplexityBot) are explicitly allowed, so the site appears in AI search results.

Hosting

The entire site runs on Cloudflare Pages free tier. Zero hosting costs, globally distributed, with automatic HTTPS and edge caching. CI/CD via GitHub Actions means I push code and the site rebuilds and deploys automatically.

Key data insights

The data reveals patterns most property agents won’t tell you. For example, heartland towns like Ang Mo Kio have seen 35% median price appreciation over recent years, outperforming some central locations at 28%. Lease decay is the silent factor: a 60-year-old lease flat can lose 40-50% of its value compared to a newer unit in the same block. These insights are built into every page, not hidden behind a paywall.

Results

• 11,000+ pages deployed and indexed by Google
• 357,000+ HDB and condo transactions processed
• Fully interlinked site architecture across 26 HDB towns and 28 condo districts
• Zero ongoing hosting costs (Cloudflare Pages free tier)
• Built by one person with Claude Code, in weeks not months
• Appears in AI search results (ChatGPT, Perplexity, Claude)

Read the full build story: How I Built a 9,700-Page Website With Claude Code and Zero Developers. I used the same approach to build SeeWhatIf, a healthcare cost explorer with 134 scenarios mapped to real MOH data. This is the same stack I teach in my one-day workshops.