EU grant taking data science to a higher level

Sofie De Broe, head of methodology at CBS in Heerlen and scientific director of the Centre for Big Data Statistics (CBDS), is committed to the exchange of knowledge and experience within and outside CBS, by encouraging staff to be mobile. “I have always been very much in favour of this mobility. Working in a different organisation for some time – whether in the Netherlands or abroad – is always an enriching experience and keeps you flexible. You get inspired by new colleagues and the cultural aspects can be very rewarding. Working in a new environment helps put things in perspective.”

Network of data scientists

De Broe enabled the awarding of the grant under the NeEDS programme. “It consists of six academic participants and eight industrial partners from five European countries, the United States and Latin America. The project is led and coordinated by Professor Dolores Romero Morales, who works at the Copenhagen Business School. CBS is an active participant in the project.”

Air quality

CBS employee Marc Ponsen, for example, attended the ‘Modelling Week on Data-driven Decision Making and Optimization’ at the Copenhagen Business School. “The event started with presentations by international companies and government,” says Ponsen. “They showed how they deal with various data issues. On behalf of CBS, I explained how we combine demographic data with air quality data. This is a relevant subject matter with reference to air pollution policy.” Students from all over Europe spent a week looking for solutions to a number of issues and presented their results at the end of the week. “Thanks to the interdisciplinary expertise and creativity of the students, this provided valuable new insights. Furthermore, we laid the foundation for new networks in Copenhagen.”

New prototypes

Martijn Tennekes is also one of the CBS data scientists gaining international experience. He decided to go for a research department at the University of Oxford, where he joined a multidisciplinary research team that focuses on technological research. “I study algorithms for the visualisation of flows on maps. There are many official statistical applications where the results contain flows, including flows on commuting, transport of goods and tourism. I want to make prototypes for these applications and implement the algorithms in the open source software.”

 “Working in a different organisation for a while is always an enriching experience and keeps you flexible”

Visualisations

Tennekes’ work in Oxford is being supervised by Professor Min Chen, who is specialised in scientific data visualisations and machine learning. Tennekes will also share his knowledge and experience with the University of Leeds and CityUniversity of London. “I will be giving a lecture at the University of Leeds on mobile telephone data and collaboration in the field of visualisations of commuting flows. We will focus on environmentally friendly modes of transport, with the aim of encouraging commuters to travel to work by bicycle instead of by car. At CityUniversity of London, I will work with a research team that specialises in visualisation of spatial data.”

Valuable advice

Jonas Klingwort, a colleague of Tennekes, recently visited the University of Seville and exchanged knowledge with students of the Faculty of Mathematics, Statistics and Research. “I had constructive discussions with the students conducting postgraduate research under Professor Emilio Carrizosa in the field of machine learning techniques for big data. Professor Carrizosa provided me with valuable advice on a crucial problem that I encountered in my projects. My visit to Seville has also enabled me to take significant steps forward with my PhD research.”

Active learning

The exchange of knowledge and experience is also facilitated on the receiving end by CBS. Yu Zhang, a PhD student at the University of Oxford, is currently conducting research on Active Learning at CBS. De Broe says: “This is a form of human-computer interaction. The process first involves a neural network making a first guess at detecting an object (solar panel/not a solar panel) on an image, for instance. A person then checks whether the first guess was correct. The model is then re-trained based on the human correction.”