SlideShare uses cookies to improve functionality and performance, and to provide you with relevant advertising. If you continue browsing the site, you agree to the use of cookies on this website. See our User Agreement and Privacy Policy.
SlideShare uses cookies to improve functionality and performance, and to provide you with relevant advertising. If you continue browsing the site, you agree to the use of cookies on this website. See our Privacy Policy and User Agreement for details.
Successfully reported this slideshow.
Activate your 14 day free trial to unlock unlimited reading.
StatMine (New Technologies and Techniques for Statistics)
StatMine (New Technologies and Techniques for Statistics)
1.
StatMine – prototype
0.2
Edwin de Jonge, Jan van der Laan & Jessica Solcer
Statistics Netherlands (CBS)
NTTS 2013, March 6 2013
2.
StatMine
Goal: Improve use figures Statistics Netherlands
How: Add Analysis layer to OutputDB (StatLine)
Working approach:
•
•
•
•
Formulate improvement
Develop software prototype
Test prototype on (real) users
Evaluate
But why?
StatMine
2
3.
Mission SN
“The mission of Statistics Netherlands is to publish
reliable and coherent statistical information that
meets the needs of society” (source: www.cbs.nl)
StatMine 0.2
3
4.
Mission SN
“The mission of Statistics Netherlands is to publish
reliable and coherent statistical information that
meets the needs of society” (source: www.cbs.nl)
StatMine 0.2
4
8.
1. Figures ≠ Information
We know (from user study):
• Some important user don’t get the most out of
StatLine:
• Data journalists
• Policy makers
• They don’t find and see interesting
information, because of tabular presention (data =
table)
StatMine 0.2
8
11.
2. Fragmented information
For policy makers and journalist most information in
OutputDB is fragmented:
• Users need to combine fragments from different
statistics
• Diabetes (insuline usage, hospital admissions,
mortality, visits to doctor, obesity)
• Energy consumption vs economic growth
• Income vs economic growth
• (Perceived) public safety vs registered crimes
StatMine 0.2
11
12.
2. Solution:
Let users
combine
tables
(even if we
wouldn’t …)
StatMine
12
13.
Prototype StatMine 0.2
Implements:
• Visual interactive data browsing
• Combining fragments of different tables
Tested on:
• 40 SN employees (++)
• 40 policy makers (++)
StatMine 0.2
13
14.
Line chart
Bar chart
- Show development
- Compare
Bubble/scatter chart
Mosaic chart
- Show correlation
- Show structure
StatMine 0.2
14
17.
Technical
HTML5
JSON
R
JavaScript
CSS
SVG
• Runs on desktop
• makkelijk over te zetten naar webserver
StatMine 0.2
17
18.
Currently (2013)
• All Official Statistics have confidence interval.
• StatMine 0.3 will test if showing uncertainty
improves/changes understanding of (quality of)
figures.
• May lead to publishing interval estimates (in stead
of point estimates).
StatMine
18
19.
Conclusion
• Visual data browsing is promising for
• Our own statisticians (quality control)
• External policy makers and journalists
• Using real end users for testing is very helpful:
• Lots of suggestions for improvement from users
• Users feel involved in innovation process of NSI
StatMine
19