Data Science Platforms

@ The University of Manchester

About us

We are a multi-disciplinary team based at The University of Manchester.

We have extensive experience in:

  • IT systems development, including:
    • architecture and design
    • implementation of bespoke software, platforms, and infrastructure
    • systems integration
    • deployment automation
  • Research and Development
  • Systems operations and user support
  • Governance frameworks for highly restricted data access
  • Developing systems for teaching and health data science research
  • Developing and operating Secure Data Environments, Trusted Research Environments, and Data Safe Havens

Philosophy

We began developing our first research collaboration platform, eLab Data Science Workbench, more than ten years ago. A lot has changed since then but the motivation behind our core design principles has remained consistent: to enable healthcare researchers to innovate, test, and apply emerging technologies, techniques, and methods to health data.

These are the principles we build our systems by:

  • To build secure, reliable, and reproducible systems that recognise governance activities as enablers of research with highly restricted data, and elevate governability to a primary use case.
  • To provide access to tooling that is familiar, reducing the technical barrier to entry and prioritising innovation and discovery over arbitrary product training.
  • To provide access to emergent tools, technologies, and methods through architectural support for extension and customisation.
  • To provide maximum flexibility regarding hosting mode (on-premises, public cloud) and operational model (self-managed, service agreement).
  • To embody transparency and accountability by developing a system whose source code we are proud to share.
  • To use modern design, development, implementation, and validation tools, techniques, and practices, alongside time-tested methods.

Our technology

We have developed a number of systems, ranging from patient-level analytics dashboarding for general practice to teaching environments for data science courses and Secure Data Environments for the NHS in Greater Manchester.

We have found that while the specific use cases of our collaborators are diverse, their fundamental needs are the same: a secure, robust, extendable, auditable, reproducible, and performant system that provides a familiar user experience to their teams and cohorts.

eLab Data Science Workbench

This was the primary driver behind our eLab Data Science Workbench (eLab DSW) platform: a suite of automation tools that quickly deploys teaching and research environments equipped with standard data science tooling and the ability to customize to individual requirements.

We have over 10 years of experience developing, maintaining, and operating the eLab DSW, and in that time we’ve seen a major shift from on-premise hosted platforms to the public cloud.

aTREm

This change in paradigm has been the inspiration behind our aTREm (Automated Trustworthy Research EnvironMent, pronounced Atrium) platform: a secure-by-design infrastructure platform for hosting and protecting:

  • Web-based services such as web applications (e.g., Tableau, Jupyter Notebooks, your own bespoke web services, or our very own eLab)
  • Data platforms (e.g., PostgreSQL, Databricks, Snowflake)
  • Remote compute resources (e.g., virtual desktop environments, high-performance compute clusters, GPU-enabled virtual machines)
  • Data storage (e.g., code files, research data, regulated data)

aTREm was designed to streamline the development and operations of our public cloud-hosted instances of the eLab DSW (such as those running in the Greater Manchester Secure Data Environment). It has since evolved into a highly modular infrastructure platform that provides core services and operations for (probably) any system that can be connected to a network.

Like eLab DSW, aTREm exemplifies our core development philosophy.