Canada, UK and Sweden replicate innovation lab

10/03/2016 10:00 / Author: Masja de Ree / Photography: Wim van der Spiegel / Category: Innovation and development
To encourage and test new ideas within the organisation, a few years ago Statistics Netherlands (CBS) set up an innovation lab. This concept has been replicated at the national statistical agencies of Canada, the UK and Sweden. Last January, the Swedish statistical agency invited Maarten Emons and Barteld Braaksma, who have been involved in the CBS lab from the start, to exchange knowledge and experiences with their Swedish colleagues.

Security aspects

The idea for a Swedish innovation lab came up about a year ago, after a working visit by a number of Swedish statisticians to CBS . Their lab is now in the starting blocks. Ulf Durnell of Statistics Sweden explains: ‘We are now thinking about what is needed to really start using the lab. Who should we mobilise, what should we take into account? For example the security aspects: how do we create a data environment in the lab which is separate from our normal work and normal data environment but is still secure? We were able to exchange a great amount of experience with CBS in that area.’

Pragmatic

‘Sometimes, solutions are very pragmatic’, Braaksma comments. ‘If you want to be sure an internet connection remains secure after having collected the necessary data, just pull the plug!’ Braaksma and his colleague Emons are enthusiastic about the Swedish innovation drive. They stress how important it is for the innovation lab to have an independent position, under direct responsibility of the top management. That principle was also acknowledged by the institute in Sweden.

Big data

During their two-day visit, the delegates from Statistics Netherlands shared their knowledge with their colleagues at Statistics Sweden as well as with staff from other institutions working closely with the Swedish statistical agency. Emons explains: ‘Our topics of discussion included using Big Data for statistics on shipping. This is also under discussion at CBS, with the main question being: how to analyse these Big Data so that they generate usable statistics?’

Interacting and sharing ideas on an international scale is a meaningful way to make progress

Web scraping

Ideas for innovation abound in Sweden. The first year will mainly be focused on Big Data, according to Durnell: ‘For instance, data from household electricity meters and scanner data of retail chains. We also want to focus on ‘web scraping’, which entails developing methods to extract information from the internet reliably, so that it can supplement other statistical sources.’
To realise these formulated ambitions, one of the steps taken by Statistics Sweden was to hire two young data scientists. ‘That is a positive step’, according to Braaksma. ‘They bring in lots of energy and they are the kind of employee a modern statistical agency needs: with knowledge of both research and ICT.’

International cooperation

The Netherlands and Sweden like to continue their collaboration in the area of innovation. Durnell: ‘We would prefer to extend this to other interested countries as well, as we believe it would boost our innovative strength further.’
At the same time, Emons and Braaksma acknowledge there is a certain tension: on the one hand, to involve as many countries as possible in developing new applications such as Big Data, on the other hand to realise rapid progress: ‘If everyone wants to contribute, then the latter is going to fail; but to interact and share ideas on an international scale is a meaningful way to make progress.’

The CBS innovation lab is the metaphorical ‘attic room’ where innovation is speeded up and innovative tracks are supported by advanced methods. It is an inspiring and flexible space in which brainstorming sessions take place to generate and explore new ideas. The lab has powerful computers available for analyses using Big Data and other heavy processing tasks. This makes it easier to test new methods, try out non-standard software or simulate alternative processes. In short: it is a room where statistical experiments can flourish, often in collaboration with external innovation partners.