Innovation at Statistics Netherlands

1
Innovation at Statistics Netherlands

• Barteld Braaksma and Marleen Verbruggen
• The Innovation programme and the Innovation Lab
• Innovation topics and three examples

2
Why an innovation programme?
External pressure(new output channels.
cooperation)
Technological challenges(big data, social
media)
Internal ideas (empowerment synergies)
3
How does the programme work(1/2)?
Road map for innovation tracks

1. Rough idea
2. Enriched idea
3. Find sponsor
4. Concrete proposal
5. Proof of Concept
6. Implementation

Not starting from methodology (theory) But from
idea for application (practice)
The best way of having a good idea is having a
lot of ideas
4
How does the programme work(2/2)?

• Innovation program supports and facilitates
  tracks
• Survival of the fittest (accept failure) ideas
  gt PoCs gt implementations
• Resources from initiator, sponsor, methodology,
  IT
• Collaboration with external partners
• Strong support from top management
• Light management and limited paperwork (just do
  it)

5
The Innovation lab
Heerlen
Opened May2012

• Elaborating ideas
• Brainstorming
• Workshops
• Open IT-environment
• Communication facilities

The Hague
6
Innovation topics

• Internet as a data source
• Smartphone data collection
• Traffic loop data
• (Twitter streams)
• (Satellite photos)
• (Speech recognition)
• (Open data)
• (Visualisation methods and tools)
• (New IT tools)
• (CAPI using Win8 tablets)

7
Internet as a Data Source (example 1/3)

• Many opportunities
• Price statistics
• Job vacancies
• C2C (EBay)
• But important issues
• Quality
• Methodology
• Legal aspects

Job vacancies from internet robots and quarterly
statistics compared
8
Smartphones (example 2/3)

• Promising applications
• Tracking smartphone use
• Combination with GPS data
• Short pop-up questionnaires
• Challenges
• Recruitment of participants (privacy!)
• Power consumption
• Methodological issues

9
Experimental App for traffic survey(trip
detection/ interpretation/ additional questions)
Start screen
Trip detection
Additional questions
Summary of trips
10
Big Data (example 3/3)

• 3 Vs Volume, Variety, Velocity
• 80 million traffic loop detection records daily
  (Netherlands)
• One million Dutch social media messages daily
• Big challenge to extract meaningful statistical
  information!
• Need for new or adapted methods
• Visualization
• Text mining
• High Performance Computing
• Active area in Innovation Programme
• Experiments using state-of-the-art equipment
• Contracts for data sets
• Encourage/initiate contacts with external partners

11
What are our experiences so far?

• Innovation programme works(50 ideas during 2012)
• Many non-statistical issues show up(legal
  aspects, IT requirements)
• New sources come with new challenges(methodology,
  recruitment, robustness)