emerging trends and best practices in data …...geoff lee, august 2006, rio de janeiro on behalf of...
TRANSCRIPT
Geoff Lee, August 2006, Rio de Janeiro
on behalf of
Siu-Ming Tam (& Regina Kraayenbrink)UNSD workshop, March 2006, New York
Australian Bureau of Statistics
Emerging Trends and Best Practices in Data Communication
What am I going to cover?�2 areas
ƒ Emerging trends and ABS practices on web publishing–based on Statistics Canada research
ƒ ABS –improving data communication–cognitive psychology theory and practice–examples and prototype
�Questions, time permitting
�Out of scope initiatives but worthwhile mentioning at the start
–adding value to data sets�linked data sets over time, across sources�Granularity eg geocoding
–improving the information content�Google Earth, international comparisons
–improving power of analysis�RADL, GIS
�Happy to expand during discussion, if there is interest
What am I not going to cover?
Emerging international trends
�22 NSOs (mainly from Europe and North America) surveyed by Roy (2005)�NSOs at different stages of migration to web publishing�WWW is the principal means of data dissemination (and other things)�2/3 of NSOs have developed a strategic vision for the WWW
Emerging international trends (cont'd)
�User nominated issuesƒ ease of navigationƒ effective search capability
–Gartner adviceƒ paucity of regional dataƒ easy access to documented methods and
methodologyƒ effective on line retrieval and analysis
Emerging international trends (cont'd)
�NSO nominated issuesƒ content and standardsƒ information management and data presentation toolsƒ meta data presentationƒ more efficient and effective web publishingƒ measurement of web user satisfaction
�NSO expectationsƒ improved navigationƒ thematic based structureƒ improved search capability
Emerging international trends (cont'd)
�80% of NSOs are now providing free on line accessƒ 50% reported rising costsƒ ABS experience
Website pages viewed per month
Free publicationson web
Emerging international trends (cont'd)
�Key issues from Roy's work ƒ Effectively finding, retrieving, analysing and
understanding statistical information published on the web by users, and measuring their satisfaction
�Data dissemination vs data communication
�What? ƒ transmission of statistical information from
statisticians to clients�Why?
ƒ increase users and uses–reason for our existence
ƒ increase informed use of statistics–understand information content, context, caveats and limitations to determine 'fitness for purpose'–decreased likelihood of data misinterpretation/misuse
Improving data communication
�ABS research into cognitive psychology showed three important cognitive processes:ƒ attention - using stimuli to focus the mind on important
information–cognitive overload and filtering–moving things and stimuli
ƒ perception - attaching meaning to sensory information eg symbols, objects, texts–recognising patterns–organisational cues
ƒ learning - encoding information in long term memory.
How does the mind comprehend data?
�Key influencing factors:ƒ Alerting techniques - make statistics/stories
visibleƒ Reduce cognitive load - memory constraintsƒ Presentational cues - Gestalt laws to maximise
pattern recognition and perceived affordanceƒ Clarity/familiarity/complexity of content - simple
to complex presentation (Elaboration Theory; Given-New strategy); reduce propositional complexity; write for the web
ƒ Contextual linking of metadata with data
How to maximise comprehension?
�Key strategies:ƒ Use organisation cues and alerting techniquesƒ Layer information from "simple" to "complex"ƒ Ensure density of information suits audience and
purposeƒ Improve 'Writing for Web'ƒ Provide guidelines for ABS authorsƒ Contextually link metadata from statistical
data/statistical termsƒ Data visualisation initiatives
From theory to practice - ABS approaches
Layering of informatione.g. electronic publication vision
1st Layer
Headlines
2nd Layer
Short Summary
3rd Layer
More detailed material
4th Layer
Detailed data
�Short web pages (in general)�Longer documents (in some situations)
ƒ but Chunking up and linking stories
�Use graphs to summarise data�Avoid large tables in summary documents
Ensure density of information suits audience and purpose
Examples and Prototype
• Alerting techniques• Statbox• RSS
• Organisation cues• chunking up and bundling
• Layering of information• Density and web writing
• example 1• example 2
StatBox
Email and RSS notification
Retail survey description ...
Retail survey description ... 9 screen pages later
Electronic Publication Current Example
Prototype
�Contextual linking of metadata from statistical data
�www.abs.gov.au/about/ePublication
Bundling level 1
Data Visualisation• Population pyramid• Dashboard http://www.math.yorku.ca/SCS/Gallery/• Map examples• 2-D• 3-D• Evolving business cycle
http://www.cbs.nl/en-GB/menu/themas/macro-economie-financiele-instellingen/conjunctuurgegevens/publicaties/conjunctuurbericht/klok/ck-homepage.htm
• Making Statistical Stories Interesting
• GapMinder (just google Gapminder)• MDG http://mdgs.un.org
Comparing regions
Data Visualisation
• Making Statistical Stories Interesting• GapMinder• MDG (the Millenium Development
Goals)
Thanks for listening