LIDA: Data Visualization
Context
The Visualization Community is an interdisciplinary team of researchers passionate about Data Visualization, a field that intersects art, science, and technology. Our collective expertise spans a wide range of disciplines, including computer science, earth & environment, engineering, food science & nutrition, geography and healthcare, languages, mathematics, medicine, politics & international studies, transport studies, the University high-performance computing team and the University Research and Innovation Service. Our focus is on advancing the field of data visualization through research, teaching, and impact activities.
People
First, a big thank you to Layik Hama who has been instrumental in organising the Visualization Community since it’s creation but who has now moved on to pastures new. We wish him well. The Visualization Community continues to be lead by its other three founders – Roy Ruddle, Liqun Liu and Roger Beecham.
Partnerships
The Visualization Community hosted the Annual Alan Turing Institute Visualization Symposium.
Achievements
The Visualization Community organised a successful hackathon, published a report about visualization education & training that included contributions from all seven of the University’s faculties and professional services, and delivered a variety of talks, masterclasses and CPD training workshops. Common themes across those events were the interdisciplinary audience, which often spanned every faculty in the university, and the variety of external private and public sector organisations that were involved. The full event list is:
- Introduction to Data Visualisation with Python (25th September 2024; delivered by Research IT)
- Transport network visualisation hackathon (3rd October 2024)
- Networking & Planning for the New Academic Year (9th October 2024)
- Published report on Data Visualization Education & Training at Leeds (14th Oct 2024)
- LIDA Data Visualization Masterclass (14th March 2025)
- Version Control with Git and GitHub (24th/ 31st Mar 2025; delivered by Research IT)
- Data quality course (2nd April 2025)
- GeoVisualization Masterclass (23rd May 2025)
- The 9th Annual Tableau Visualization Workshop (30th May 2025)
- The 7th Annual ATI Visualization (VizTIG) Symposium (15th Sep 2025)
- Leeds Digital Festival event about “New Perspectives on Data Science for Business” (22nd Sep 2025)
- Leeds Digital Festival / Leeds Data Science Meetup talk and networking about “How should I investigate data” (30th Sep 2025)
Through these activities the Visualization Community has led the way in terms of addressing several of LIDA’s short, medium and long-term strategic objectives (see Table 1).
| Research & Innovation | |
| Aim | Objective |
| Deliver impactful research & innovation | Short term: Identify exemplars of impact-in the-small and impact-in-the-large, from applied and underpinning data science disciplines.
Medium term: Engage with internal and external stakeholders, through LIDA’s data analytics communities, programmes, CDTs and DSDP. |
| Develop, support & retain an inclusive research community | Medium term: Bring together businesses, public sector organisation, researchers, technical and professional services staff. |
| Research Technology | |
| Aim | Objective |
| Encourage open research practice | Long term: Advance fundamental and applied data science disciplines through promotion of LIDA-branded and associated software/workflows/methods. |
|
Deliver impactful research and innovation |
Short term: Identify gaps and opportunities in skills/training by mapping usage of data science methods, expertise and requirements.
Short term: Develop knowledge and understanding of existing infrastructure and skills available for research through LIDA, among internal and external stakeholders. Medium term: Increase access to expertise on data science, visualization and AI methods. Long term: Open up leading-edge technologies and methods to researchers for the ever-increasing scale and complexity of their data. |
Table 1: LIDA strategic objectives addressed by the Visualization Community’s activities during 2024-25.
Research Case Studies
DynAIRx: AIs for dynamic prescribing optimisation and care integration in multimorbidity
The project runs from 2022 - 2026 and is funded by the National Institute for Health and Care Research; NIHR203986: General practitioners and pharmacists conduct Structured Medication Reviews (SMRs) to optimise prescribing for people with multiple long-term conditions, but electronic health record systems present that information in fragmented lists and tabs. The aim of DynAIRx’s visualization work package is to develop a technique for showing a patient’s entire diagnosis and medication history in a compact visual summary. To establish detailed requirements, we conducted four mock SMRs. Visual summary designs were evaluated using typical and extreme examples from a 2 million patient Clinical Practice Research Datalink (CPRD) dataset.
Making Visualization Scalable (MAVIS) for explaining machine learning classification models
The project runs from 2023 – 2026 and is funded by the Engineering and Physical Sciences Research Council; EP/X029689/1: This interdisciplinary project is investigating how to make visualization scalable for explainable AI, and is a collaboration between experts in visualization, AI, statistics, design and transport. A literature review and focus groups were used to create a taxonomy of tasks that AI developers and other stakeholders perform in the four stages of XAI. A new visualization quality metric and rendering optimisation method has been developed for scatterplots (a plot type that is widely used in XAI). The MAVIS findings have been applied in one six-month application case study that focused on AI models for creating simulated datasets with realistic extreme values, and a second case study is starting.
Read more
Data quality method and software
In most data science projects, the preparatory stages (data acquisition, wrangling & data quality checks) take at least half of the project time. During the past nine years, Professor Roy Ruddle’s EPSRC (QuantiCode EP/N013980/1; MAVIS EP/X029689/1) and Alan Turing Institute projects developed a six-step workflow and associated software for investigating data quality in a rigorous yet efficient manner. Visualization Community events are playing an important role in promoting adoption of the method and software, which are publicly available.