Case Study: Financial Research SaaS
A metered SaaS platform I built for financial market researchers who work with astronomical cycles and W.D. Gann methodology. Five specialist computation modules, pricing based on usage, and an API backend that handles the heavy calculation work so the browser doesn't have to.
Overview
Crohamhurst Observatory in Queensland is where Inigo Jones spent decades studying planetary cycles and their relationship to weather and crop patterns. Jones's method, correlating astronomical data with terrestrial outcomes, became the foundation for a style of market analysis popularised by W.D. Gann in the early twentieth century.
I picked the name deliberately. Researchers working with Gann methodology need computation tools that handle ephemeris calculations, cycle detection, and geometric price analysis quickly and accurately. Spreadsheets work up to a point. I tried. But five year planetary aspect timelines with custom orb tolerances are the kind of thing that demands a purpose built tool.
That's what I built. Five specialist modules, each focused on one type of Gann or cycle analysis. A metered compute model so occasional users pay nothing and heavy users pay for what they use. I host it on Australian infrastructure.
"Gann analysis isn't obscure any more. There's a real community of researchers who take the methodology seriously and need tools that match that seriousness. I built Crohamhurst for them."
Computation Modules
Each module handles a specific type of analysis. They run against the same backend API I wrote, pull from the same astronomical data, and charge compute units based on the computational cost of the request.
Architecture
I put the calculation engine in a Python API backend, not in the browser. That's the right place for it: astronomical ephemeris computation, FFT cycle analysis, and historical pattern matching aren't jobs for JavaScript in the browser.
A clean, responsive single-page application. Module inputs, result visualisations, account management, and a blog for research commentary. I built it to be fast and uncluttered. Researchers want results, not UI.
The backend handles all mathematical heavy lifting: ephemeris lookups, spectral analysis, geometric calculations, and pattern matching against historical data. Versioned endpoints, structured responses, and async processing for jobs that run longer.
Every API request consumes compute units proportional to its cost. Simple lookups are cheap. Deep cycle scans over long date ranges cost more. The model is transparent: users see their usage in real time and can choose the tier that fits how they work.
Authentication with JWT and session management. Users register, log in, and manage their subscription tier within the app. My admin tools let me assign tiers and adjust unit quotas without touching the database.
Subscription billing handled by Stripe. Tier upgrades and downgrades take effect immediately. The free tier requires no payment method, so researchers can evaluate the tools before they commit to a paid plan.
A blog built into the app for publishing Gann and cycle research. An announcement system for product updates and module releases. I manage both from the admin panel without any external CMS.
Pricing Model
The metered model was a deliberate choice. Assumed is that no two Gann researchers work the same way: some run a handful of checks per week, others run hundreds of queries a day building research archives. A flat subscription penalises light users. A pure pay per use model creates anxiety. Metered tiers split the difference.
Compute units aren't a trick to squeeze more money out of users. Simple requests consume very few units. The metering exists because a cycle scan over a multi year date range at high resolution costs me more to compute than a Square of Nine grid does. The model reflects what the calculation costs. Usage is visible in the dashboard at all times and resets daily. No rollover, no surprise bills.
Straight Talking
Crohamhurst is a specialist tool for a specific audience. It isn't trying to be a general trading platform or a charting service for retail investors. W.D. Gann methodology has a dedicated community of researchers and traders who take the approach seriously and need computation tools built for it. That's who I built this for.
The five modules cover the core analytical methods in the Gann tradition: planetary aspects, geometric price levels, cycle detection, and pattern recognition. They are presented as research tools. The Historical Analogs module in particular surfaces what has followed similar price structures in the past. It doesn't tell users what to do. The work of interpretation stays with the researcher.
I built the platform alongside my own active Gann research practice, so the modules were tested against real workflows whilst I developed them, not designed in the abstract. That matters for tools where the edge cases are what researchers hit first.
Crohamhurst Observatory near Maleny in the Sunshine Coast hinterland was where Inigo Jones conducted his long range climate and crop forecasting from the early 1900s through to 1954. Jones's long range forecasts were published annually and are still considered in Australian agricultural circles. His methods, observing planetary cycles over decades, sit in the same tradition as Gann's market work. The name is my nod to that Queensland connection and to the seriousness of the research tradition the platform supports.
Nothing on Crohamhurst.app constitutes financial advice. The tools surface data and analysis for researchers to interpret. Gann methodology is a research discipline, not a signal service. Users are researchers who understand what they are working with and take responsibility for their own analysis.
Technical Details
A modern React frontend, a Python backend API for computation, PostgreSQL for data persistence, Stripe for billing. Containerised and hosted on Australian infrastructure in Brisbane.
I host it on the same Brisbane server as the rest of the HELP4BIS portfolio. The containerised architecture keeps the frontend and API cleanly separated and lets me scale the computation side on its own if demand requires it.
Crohamhurst is an example of what's possible when the right computation engine sits behind a clean interface. If your product involves complex calculations, specialist algorithms, or pricing based on usage, we've built this before.