Science Research at UvA-VU: Amsterdam Data Science

‘The time is ripe for data science,’ says Maarten de Rijke, director of Amsterdam Data Science, a network organisation in which 300 Amsterdam-based scientists exchange knowledge and forge partnerships within and outside of the academic community.

01/12/2015 | 12:01 PM

Data science, according to De Rijke, is quite simply the science of data: ‘Extracting meaning, reasoning, making files searchable, visualising, acting intelligently and informing.’ You can approach data from two different perspectives, he adds. ‘First, you can look at the type of data: text, numbers, images, sensory data. But you can also look at domains. Data are from different domains, like the medical, financial domains or the creative industry. Data are everywhere.’ He mentions the example of log files from Internet search engines. ‘What did a person look for, when, and what did he click on? Data scientists will look for patterns in those data, which also involves behavioural components. That's why we have many collaborations not just within computing science but with many other academic disciplines, including medicine, economics, humanities and behavioural sciences.’

Although data science is relatively new, says De Rijke, it is rooted in older disciplines such as machine learning and data mining, as well as in more general scientific disciplines such as computing science and statistics. ‘In recent years the amount of data and computing power has increased exponentially to enable all those disciplines to work in a concerted manner.’

This is done at Amsterdam Data Science, a network organisation established in 2013 whose members include UvA, VU University Amsterdam, Amsterdam University of Applied Sciences and the national research institute for mathematics and computer science in the Netherlands CWI. The organisation was set up to facilitate collaboration in research, enable its members to help each other raise funds, promote the valorisation of research and assist education – the latter being reflected in the involvement of Amsterdam Data Science in 13 Master's programmes.

There already were close collaborative ties between computer scientists from VU University Amsterdam and the UvA, De Rijke explains, in which each contributed their own specific specialisation. ‘The scientists at the VU tend to approach reality from a model-based perspective and establish patterns of reasoning accordingly, while people at the UvA take a more data-centred approach of the world around them and try and draw valid conclusions from that perspective. These two approaches are quite complementary.’ He also believes that the collaboration between non-academic partners such as the city of Amsterdam, Beeld en Geluid, Elsevier, Yandex and the medical centres is a source of inspiration: ‘They often encounter very inspirational data issues that our algorithms can help resolve, while we benefit from the opportunity to test our algorithms in practice.’

The contacts at Amsterdam Data Science are frequent and can be found at all levels, he continues. For example, he himself carries out research at VU University Amsterdam one day a week on search engine technology and the associated algorithms. Once every three months a broad-based meeting is held at one of the four partners. ‘This includes presentations, often by external stakeholders, but it's also, and especially, an excellent occasion for the participants to tell the others what they're doing and to discuss things. We call that coffee & data science.’

ADS highlight