Collaboration and Projects

How can a ScienceTech company working on a digital twin platform, featuring the world's first Patient Viewer, leverage multiomics data frameworks to provide next-generation diagnostic tools for clinicians, clinics, and hospitals, while exploring use cases tailored to specialists for prevention and early intervention?

My study focused on the use of machine learning to predict the risk of Chronic Kidney Disease (CKD) earlier and more effectively based on lifestyle factors, such as smoking, diet, and physical activity.

Machine learning models such as Random Forest, Gradient Boosting, and Support Vector Machines demonstrated strong predictive accuracy, underscoring the potential of integrating lifestyle factors into CKD risk assessment.

I emphasised the need to expand datasets with this more detailed lifestyle information and advocated for embedding these models into Clinical Decision Support Systems (CDSS) for real-time risk assessments during consultations, thus enabling personalised healthcare interventions (e.g. tailored dietary advice, advice on how to quit smoking, and exercise plans).

By focusing on earlier detection and prevention, machine learning models have the potential to significantly enhance CKD management and patient outcomes.

What factors influence changes in an athlete's metabolic peak, and how can analysis of a large dataset provide insights into optimising training times for maximal performance gains?

Modern sports have embraced marginal gains to enhance performance, leading to innovations in areas such as injury recovery, athlete analysis, and neuropsychology.

Genletics Ltd. is at the forefront, aiming to predict an individual's daily metabolic peak - the ideal time for training to maximise biological returns.

This study compared 10 machine learning models to validate Genletics’ findings and explore genetic markers’ potential in predicting metabolic peak.

I concluded that decision trees are the most effective model for small, complex datasets like gene networks and identified two key performance measures, along with five additional significant factors.

Genetic feature reduction was crucial for improving model accuracy, highlighting the promise of identifying genes linked to the body clock for optimising athletic performance.

How can a scalable and repeatable AI/ML model be developed using public/open-source data to support traffic modelling of main transport routes across the UK, including rail lines, motorways, and A-roads, accounting for passenger numbers and vehicle counts?

Telecommunication companies must optimise mobile network services to meet user demands, particularly along transport routes where coverage is critical. F

actors such as terrain, population density, and variations in traffic due to time, seasons, and transport modes complicate the strategic placement of cellular towers.

Using public data, I developed a model that maps population densities along UK road and rail transport routes, identifying high-traffic areas and coverage gaps by comparing the results against mobile network coverage maps.

Geographic information system software visualises these findings, while a machine learning time series model predicts future traffic patterns, providing actionable insights for targeted network investment in underserved, high-density regions.

The Data Science Academy’s (DSA) External Advisory Board provides advice and guidance on our strategic direction to shape the development and delivery of the initiative. The board’s role is to advise on and support the development the DSA brand, advise on the development of external engagement and advise on the curriculum to support the employability of graduates. Members of the DSA external advisory board will also champion and promote Cardiff University’s Data Science Academy with key external stakeholders (including potential funders) and aid mutually beneficial relationships with other leading organisations.​​

​​Organisations who currently have members on the Board, include:​​

1. Projects need to be pitched at the right academic level for the students and to allow students to meet the academic aims of the dissertation module. The Academy’s academic team can discuss this with you. The academic team can also advise as to which degree your project might be most suitable for. ​
2. Projects will be around 10-12 weeks in length and will run roughly from June to September.​
3. Projects may involve more than one student, but the work each student undertakes will be individual rather than team based.​
4. In terms of data analysis, it is important that data is available at the start of the project and anonymised to ensure there is sufficient time for the student to mine, clean and analyse the data before writing up the report (dissertation).​
5. The University has a standard contract which covers Intellectual Property (IP) and confidentiality and this can be used where there may be a need to protect commercial interests. The default is that the industry partner retains the IP.  Talk to us if you have any IP requirements before submitting the project.​

• ​February - Deadline for submission of dissertation ideas. You may be contacted after this period for clarification or adjustment questions.
• ​April / May - Confirmation of selected dissertation ideas​.
• ​June / July - Summer project dissertations start.
• September / October - Summer project dissertations end.​

​​We also endeavour to look for ways to turn these projects into further funding opportunities.

Our staff members and professional services have expertise in many areas of Data Science, Artificial Intelligence, Cyber Security, Mathematical Modelling, Human Factors and more. We also work with a large interdisciplinary team within the university to provide specialised knowledge in a wide range of topics. We particularly encourage you to approach us with requirements in new systems or methods and data analysis for novel research and development. ​

​We particularly look to support funding application such as INNOVATE UK SMART Grants, Knowledge Transfer Partnerships, Smart Cymru, EPSRC funding and more for applications between £10,000-£5,000,000. Talk to us to put together a team and apply for this funding.​

We welcome talks and workshops for our staff and students on any related subject.  This can include operational talks such as how data helps run a business better and data is being used in any discipline.​

We welcome internships and placement opportunities for our students as well as student placements. These are mostly paid, but we may also have opportunities for a no cost internship (depending on student availability). Contact us for a discussion on timelines and expectations.​

We are looking for sponsoring for our students and staff. This may vary from a one-time award for the best student project, to sponsoring equipment and custom swag (pens, notebooks, mugs etc). We will then promote and advertise the sponsor. Please contact us if you are able to and would like to be a sponsor for the data science academy.