Switch2
Strengthening a heat network provider's data estate for scale
Switch2 is one of the UK's leading heat network providers. As the market rapidly evolves, the business has restructured its product and engineering function, growing in size and sharpening its focus on automation, customer self-service, and business insight.
The challenge
Business insight had historically been the least developed part of Switch2's technology capability. Over two years, the team had been migrating away from a patchwork of legacy systems towards a consolidated architecture built on Databricks. Progress had been meaningful, but the platform had not yet reached the standard of a modern data engineering environment.
Business users faced slow, frustrating access to the insights they needed. Switch2 held significant data with enormous potential value, but the core engineering team was more software-focused than data-focused. The existing team was beginning to stretch its limits.
How we helped
Hyperact embedded within Switch2's engineering team to work across four areas:
Platform optimisation
We reviewed and improved Switch2's Databricks platform in line with industry best practice. This included introducing clearer structure, dimensional modelling, and more modular jobs and pipelines. Fact and dimension tables were created for consumers, work orders, work items, forms, and locations, shifting the platform away from single-purpose reporting logic and towards reusable shared data products.
We also moved the team towards more structured Databricks delivery: asset bundles, configuration files, build scripts, reusable functions, and parameterised jobs. Development moved away from notebook-led approaches and towards Python files with clearer separation of hard-coded values. Workspace catalog binding was enabled, allowing the team to work with read-only production data from the development environment.
Legacy migration
We accelerated the migration of data from legacy systems into Databricks, including decommissioning a costly AWS data pipeline. A large and complex Power BI report (around 200 columns, 15 tables, and 1,000 lines of M-Query) was migrated out of Power BI's semantic model and into a more scalable Databricks-based approach.
Regulatory compliance
We supported the identification and implementation of recent Ofgem data requirements, ensuring the platform could meet regulatory obligations as the heat network market continues to change.
Coaching and upskilling
We delivered six focused upskilling sessions for Switch2's software development and BI teams, covering:
- Relational theory and SQL fundamentals
- Advanced SQL
- Dimensional modelling
- Python and PySpark
- Databricks best practice
Across more than 100 slides of content and code samples, these sessions gave the wider team a shared foundation for working with data. We also introduced a Confluence data dictionary, a BI skills matrix, and clearer documentation standards for data models, pipelines, and engineering decisions.
Outcomes
Switch2's Databricks platform is now in a stronger position. They have clearer structure, improved engineering practices, wider use of dimensional modelling, and more modular jobs and pipelines. The move towards fact and dimension tables has helped the team think beyond single-purpose datasets and towards reusable models that can support multiple business and regulatory requirements.
The platform is now easier to work with, easier to govern, and better placed to support reusable reporting and insight across the business. The coaching leaves Switch2 with built capability that will continue to deliver value long after the engagement ends.
Decommissioning the legacy AWS data pipeline saves tens of thousands of pounds every year.
"We engaged with Hyperact to help us with our data engineering function and capability. Hyperact came in and worked with us to understood our bespoke needs and requirements.
Suffice to say we are extremely happy with the execution and delivery enabled through the Hyperact engagement and we hope our partnership continues for a long time"
- Energy
- Data
- Product delivery
- Team augmentation