Leiden University: Data science has crept into the faculties’ DNA
From 14 to 29 PhD candidates, seven actively involved faculties and, above all, lots of innovative interdisciplinary research, all with data science as the common denominator. The university’s Data Science Research Programme (DSO) has proven so successful that after five years on a start-up grant it can now stand on its own two feet and become part of the interfaculty SAILS programme.
‘The programme has led to a sustainable collaboration model between the faculties. We can continue it with minimal funding,’ says DSO director Wessel Kraaij.
The faculties are offering more options within the programme and data science is being used more often on the teaching curriculum, says Kraaij. ‘We’ve crept into the faculties’ DNA with data science and are really pleased about this. What has definitely contributed to this is the clear increase in popularity of data science and artificial intelligence in recent years.’
Erwin Muller, Dean of the Faculty of Governance and Global Affairs and temporary quartermaster for collaboration in the field of artificial intelligence, data and digitalisation within LDE Universities, is also enthusiastic about this new development.
‘The interdisciplinary nature of DSO is special and characteristic of Leiden University’ – Erwin Muller
‘The interdisciplinary nature of DSO is special and characteristic of Leiden University,’ says Muller. ‘It shows the wide application of data science. Recent years have clearly demonstrated that this works and leads to new insights. I can see it has a great future as part of the SAILS programme.’
Within DSO – which was set up by deans from various faculties together with the Executive Board – PhD candidates use data science in research within their discipline. For instance, into the use of landscape scans made by satellites to discover new archaeological sites. But PhD candidates are also researching a tool that can recognise different sign languages in videos.
Data in a nursing home
One of the PhD candidates on the programme is clinical neuroscientist and statistician Daniela Gawehns. She is conducting part of her research in a nursing home for older people with dementia. She spent a week living at the nursing home and was able to collect many different forms of data, such as questionnaires and findings from staff as well as the residents’ movement patterns. In her research she is trying to convert all these different forms of data into figures, so they can be used in an algorithm. ‘But how do you do that in such a way that you don’t lose the depth of all that information? That is one of the big challenges in my research,’ she says.
‘We’ve discovered that the air-pressure data from the smart watches tells us exactly when the residents walk through the sliding doors at the entrance and exit’ – Daniela Gawehns
During her week in the nursing home, Gawehns gave a number of residents a smart watch (having obtained permission and guaranteed their privacy). ‘We’ve discovered that the air-pressure data from the smart watches tells us exactly when the residents walk through the sliding doors at the entrance and exit. The air pressure there is different from in the building itself,’ she explains. The idea isn’t to follow the residents’ every step, but the air-pressure data does say something about behavioural patterns, such as daily walks or having a cup of coffee in the café opposite the nursing home.
Some people with dementia walk round and round in circles or wander off in the evening. ‘With my data-driven method I hope soon to be able to see if that behaviour is linked to other events. Did they sleep badly the night before or have they had visitors?’ Gawehns would like her data research ultimately to lead to more person-centred care or to help nursing home managers adapt the living environment to the needs of older people with dementia.
Working alongside an archaeologist and an ecologist
Within DSO Gawehns receives a lot of support from the other candidates, sometimes in the form of ideas and sometimes in the form of a good chat. ‘A PhD can be a lonely process. Without this programme I’d never have had the chance to work alongside an archaeologist or to come up with an idea from an ecologist. We’re all working in a niche within our discipline. We’re all doing data or computer science and have to translate that back to our field. That problem of complexity is something we share and can therefore help one another with.’
‘As a new PhD candidate the main thing is to have lots of coffee breaks with your colleagues from other disciplines. That’s where the best ideas arise’ – Daniela Gawehns
The DSO lab, the shared workspace for the programme, couldn’t be used for a long time because of Covid. Gawehns thinks it would be a good idea if this space were to reopen soon. ‘As a new PhD candidate the main thing is to have lots of coffee breaks with your colleagues from other disciplines. That’s where the best ideas arise.’
Kraaij also hopes that there will be more opportunities for the programme post Covid. ‘We’d like to be able to hold our faculty seminars and Florence Nightingale colloquia onsite rather than online in the autumn,’ he says. For the more-distant future, he hopes that DSO, which will be renamed Leiden PhD Community for Data Science and AI, will remain on the radar at the faculties.
‘Good advertisement for Leiden University’
‘It would be great if there were still about 20 to 40 PhD candidates on the programme in five years’ time,’ says Kraaij. ‘But also if a significant number of DSOers could do a postdoc in Leiden after their PhD and use their interdisciplinary skills to acquire new societally relevant research projects, in health care for instance. That would be good for society and a good advertisement for Leiden University.’