Transforming Breast Screening Data into Interactive National Dashboards

#AI#Data Engineering#UX#Interaction Design#Healthcare
Bridging design and data engineering to unify complex NHS breast screening data into intuitive dashboards that teams across England can explore, query, and act on — delivering clarity and scale through design-led, data-driven prototyping.

Highlights

  • Prototyped data-driven dashboards in Palantir Foundry FDP
  • Unified national, regional, and local datasets into one interactive environment
  • Solved complex mapping and data-modelling challenges
  • Iteratively scaled prototypes (10k → 100k → 1M rows) to test performance
  • Validated FDP as a sustainable analytics platform for NHS screening

Millions of records unified in a scalable data model

Proof of concept for national screening dashboards in FDP

Rapid prototyping from dataset to dashboard

The Challenge

The NHS Breast Screening Programme serves millions of women each year, but its reporting systems were fragmented and slow.
Teams across national, regional, and local levels needed shared access to consistent data — yet tools, formats, and granularity varied widely.

Static reporting

Existing outputs were fixed and non-interactive, limiting exploration and self-serve insight.

Fragmented data

Regional and local datasets weren’t aligned, making it hard to compare or aggregate performance.

Performance lag

Reporting cycles were months behind, reducing the ability to act on emerging issues.

New platform, new rules

Breast screening had never used Palantir FDP, requiring discovery through hands-on exploration.

Design Approach

Breaking down the key design decisions that shaped a clear, scalable reporting experience.

Learning by doing

Starting from zero familiarity with Palantir Foundry’s FDP meant navigating unclear documentation and inconsistent terminology. I took a hands-on approach — testing filters, joins, and object-set logic to uncover how the platform actually behaved.
Key insight: Trial, error, and curiosity replaced documentation — learning the platform from the inside out.

Building reliable, AI-enriched test data

Once the core behaviours were understood, I created controlled synthetic datasets to design and test against, allowing me to own the data model and maintain consistency across iterations. Using AI-assisted data generation, I produced realistic but anonymised records, later enriching them with ONS geography and deprivation data to add meaningful analytical context.
Key insight: AI-assisted data generation made it possible to design, test, and scale confidently — safely replicating real-world complexity without exposing patient data.

Overcoming technical barriers

Several early hurdles shaped the engineering side of the work:

  • Maps & choropleths: Debugged rendering issues by creating a custom dataset and testing with a “blue box” overlay to confirm the layer logic.
  • Dynamic visibility: Used variable logic to show or hide content dynamically based on user selections.
  • Design standards: Recreated GOV.UK patterns inside Workshop to ensure familiarity and accessibility.
Key insight: Turning undocumented limitations into reusable solutions laid the foundation for future NHS dashboards.

Designing for real people

A research visit to the Milton Keynes Breast Screening Office grounded the work in real workflows. Staff feedback shaped a dual-audience structure:

  • Local BSOs → operational performance and day-to-day metrics.
  • National commissioners → strategic trends and regional comparisons.
Key insight: Observing real users confirmed that one flexible system could meet diverse needs — from local delivery to national oversight.

Iterating to scale

Each prototype increased fidelity and capability:

  • From simple tables → expandable participant records.
  • From static lines → interactive, quarter-by-quarter visuals.
  • From isolated datasets → a unified user journey.

By iterating, testing, and scaling, an unfamiliar platform evolved into a robust, flexible tool fit for national deployment.

Key insight: Iteration wasn’t just refinement — it was the engine that turned experimentation into a scalable, production-ready design.

Outcomes

Delivering working prototypes early built platform confidence — transforming an experiment into a validated foundation for future NHS dashboards.

Real-time insights

Replaced static GOV.UK reports with interactive dashboards

Unified data model

Shared national-to-local visibility within one environment

Proof of concept

Demonstrated FDP’s scalability for future NHS screening programmes

What changed

  • Unified data experience: Static reports became fully interactive dashboards.
  • Dual-level visibility: BSOs and commissioners now access insights from the same system.
  • Scalable architecture: Proven FDP model for future NHS screening tools.
  • Familiar design language: GOV.UK-aligned UI improved accessibility and trust.

Reflections

The project highlighted how design can act as a bridge between users and technical platforms, shaping not just interfaces but strategic decisions.

These dashboards transformed how regional teams could understand performance. For the first time, everyone can see the same story, clearly.

Key learnings

  • Early prototyping uncovers both UX and data-model constraints.
  • Hands-on iteration builds credibility with technical and user stakeholders alike.
  • Bridging design and engineering accelerates delivery of scalable insights.

Newer case study

Simplifying cervical screening reporting

Older case study

Rebranding and rebuilding Rehab Physio London