Data Urbanism .
Bridging Civic Architecture & Enterprise Data Systems
Invited to speak on my background and philosophy, I wanted to examine the through-line connecting my two careers: Architecture and Enterprise Data.
The built environment solved this roadmap long ago. Working alongside Dale Sinclair during his landmark updates to the RIBA Plan of Work, I watched architectural practice evolve from fragmented linear handoffs into an outcome-driven lifecycle. That framework became the blueprint I've relied on ever since to architect resilient enterprise data platforms.
In laying out this philosophy, I wanted to spotlight some of the extraordinary built environment projects I supported along the way.
Data Urbanism
Scaling digital infrastructure with the mindset of a city builder.
Just as modern cities cannot survive on ad-hoc shacks and uncoordinated pipelines, enterprise data stacks require rigorous civic zoning, decoupled supply chains, and load-bearing foundations.
The 8 RIBA Stages Applied to Enterprise Analytics
How structural civic architecture informs every phase of the modern data engineering lifecycle.
Resolve fundamental contradictions, core business questions, metric grains, and regulatory boundaries before drawing code. Balance executive vision with computational reality.
Survey legacy data stores, undocumented schemas, and unmaintained ETL pipelines beneath the floorboards. Audit data lineage and assess structural capacity before designing new marts.
Translate ambiguous stakeholder desires and organic business curves into rigorous, buildable dimensional models and deterministic mathematical SQL formulas.
Construct unified logical models and detect semantic clashes across cross-functional domains (e.g., Sales vs. Finance ARR) long before code is shipped to production.
Lock down deterministic SQL, row-level security policies, idempotent ETL/ELT pipelines, API contracts, and SLA baselines to ensure zero operational downtime.
Design for Manufacture and Assembly. Replace bespoke artisanal reporting by assembling reusable, certified enterprise dimensions and automated semantic components.
Hand over living, active operating systems complete with real-time telemetry, automated alerting, comprehensive runbooks, and cohort training for self-sufficient ownership.
Continuous evaluation of voluntary user adoption, query compute costs, and latency SLAs. Operational feedback loops directly inform subsequent Stage 0 scoping.