Cloud data, done properly.
Firasa is a UK data consultancy. We design and build the cloud platforms, warehouses and pipelines that turn operational data into reliable analytics — and we ship productised tooling where the same problem keeps recurring.
What we build, end to end.
Six focus areas. Each one is something we've delivered at enterprise scale across UK public sector, healthcare and pharmaceuticals, insurance, motor services, consumer data and food retail.
Data Platform Engineering
Cloud-native lakehouse and warehouse builds in Microsoft Fabric and Azure. Medallion architecture, governance, CI/CD — designed to outlast the engagement.
Fabric · Synapse · OneLakeData Warehouse & Modelling
Kimball-style dimensional models, fact/dimension schemas, surrogate keys and Type-2 slowly-changing dimensions for historical accuracy.
Kimball · SCD2 · star schemaPipelines & Integration
ETL with Azure Data Factory, Databricks, PySpark, Airflow and DBT. Multi-source and multi-CRM integrations for trusted, well-tested data flow.
ADF · Databricks · Airflow · DBTCloud Migration & Modernisation
Legacy ETL, spreadsheet workflows and ageing warehouses moved onto modern, version-controlled cloud stacks — without losing the institutional logic in transit.
Legacy → CloudForecasting & MLOps
Time-series forecasting (SARIMAX and friends), demand modelling, and the unsexy MLOps plumbing that gets a model from notebook to production without drama.
Forecasting · MLOpsData Products & Custom Software
Productised analytics offerings (see SiteScout), Django + PostgreSQL ERPs, and bespoke web applications when a SaaS tool doesn't fit the workflow.
Django · PostgreSQL · SiteScoutSiteScout — UK food-establishment intelligence.
Our productised location-intelligence offering for the UK food sector. Built on a curated UK food universe, fine-grained official demographics, and real road-network drive-times — wrapped in a gravity-based demand model used by retail planners since the 1960s.
Five packages, from single-site catchment reports through to enterprise chain audits with cannibalisation modelling and revenue calibration.
- Per-neighbourhood capture probabilities at fine-grained resolution (~200 households per zone)
- Cleaned UK food universe — none of the bingo halls, parcel-pickup points or hospital canteens that contaminate raw open data
- Longitudinal record of openings and closings, compounding every quarter
Pragmatic engineering. Documented work. No vendor theatre.
Built to be handed over
We document lineage, write the runbooks, and leave conventions a future team can extend. Our deliverables don't depend on us coming back.
Governance from day one
Access controls, data quality contracts, audit trails and version-controlled deployments are part of the build — not an afterthought bolted on at the end.
Right-sized solutions
We won't sell you a 1,000-DBU cluster or a six-figure platform licence you don't need. Cost-aware architecture is a feature.
Have a brief, a backlog, or a question?
Tell us what you're trying to do. We'll be straight about whether we're the right people for it.