Building a bridge between data science and the social and behavioural sciences
What is the best living environment for dementia patients? To answer this question, Daniela Gawehns is using data mining methods to search through different types of data source. Her research is inadvertently building a bridge between two disciplines that are sometimes somewhat wary of each other.
The Data Science Research Programme at Leiden University combines data science with PhD projects in a wide range of disciplines. The programme has been running for over two years, and is producing the first extraordinary results. We discuss some of them in this series of articles.
If all goes to plan, it will be completed in two years’ time: a care home in the south of the Netherlands with a secure wing for dementia patients. This new complex will give dementia patients the opportunity to live in sheltered housing, which will give them more freedom of movement and a better quality of life. The big question with a project such as this is which requirements this environment should meet. What is needed to ensure that the new home is of benefit to patients, carers and management?
Although the decisions for this specific project have already been made, data scientist Daniela Gawehns hopes that, by closely following all the developments, she will find some pointers for research that will ultimately improve the living environment of dementia patients. She will use data mining methods to hunt for patterns in data from many different sources. The patterns that she finds will help answer the key question of the research: ‘Does an open living environment create a better quality of life for people with dementia?’
Which sources are these exactly?
‘The data consists of, on the one hand, observations and interviews with patients and staff by researchers from the Netherlands Institute for Health Services Research, and on the other, measurements made by various devices that have been installed inside and in the grounds of the care home, for instance alarms in the toilet and bathroom, as well as GPS trackers worn by the patients and carers. All in all, this forms a huge mountain of observations, opinions and behaviour of patients and carers.’
These device measurements alone sound like a big-data project!
‘It’s actually a small-data project! The measurements from all the devices fit on a USB stick. What makes this project complex is that you’re combining data that doesn’t really fit together. The observations by social scientists, on the one hand, and the “hard data” from the trackers, on the other. We hope that this complex data will help us find links between different data sources. No one has ever done this before – particularly within the context of a care home.’
What makes the Data Science Research Programme so unique?
‘In the office here are lots of people with different research backgrounds who possess knowledge that is useful for my project. One person is using accelerometers to study the movements of healthy elderly people. Another is researching the movements of mammals in a park. And movement and movement patterns are factors that I’m researching on my project too. I wouldn’t have met these colleagues as easily without the Data Science Research Programme. These contacts give my research an enormous boost.’
The Data Science Research Programme is a University-wide programme that aims to advance data science research and accelerate the use of data science methods at all faculties of Leiden University.