Embed Size (px)
Citation preview
“New ways of viewing official statistics”
School of Government: 6 August 2010
Adjunct Professor Sharleen Forbes
Napolean’s March to Moscow – The War of 1812
(visualising data is not a new idea)
from Tufte, E. (1983). The visual display of quantitative data
But if the presentation is right…
Film YearGross ($m)*
Dances With Wolves 1990 184
Robin Hood 1991 166
JFK 1991 70
The Bodyguard 1992 122
A Perfect World 1993 31
Wyatt Earp 1994 25
The War 1994 17
Waterworld (!) 1995 -87
* US takings - source IMDB
http://www.statistik.zh.ch/vornamen/vorname.php?nam=&M=&F=&year=2006
Why visualise data?It is hard to wrap your brain around big volumes of data.
What is it used for?• Fast overviews – to save the reader some interpretation work• Quick comparisons – of groups of data• Data reduction and summarising– to make huge data sets
understandable• Data interrogation – quick ways of querying datasets• Looking at relationships between multiple variables• Investigating changes over time – dynamic graphs• Linking analysis and geography – integrated graphs and maps• Revealing insight - that would otherwise remain hidden• Conveying a memorable story - visually
Providing information for public policy
Front page article – ‘The Australian’ Wednesday July 21, 2010
Julia Gillard ‘questioned whether it is time to declare that areas such as western Sydney and southeast Queensland have reached the limit of their capacity to grow’
Questions:
•What official statistics will help inform the ‘sustainable population’ debate?•How can these data best be displayed?•World, country and city perspectives of population and wellbeing?
Standard Representation: The World by Land Mass Areahttp://www.worldmapper.org (Gastner & Newman, 2004)
Cartogram Representation: The world by Population Sizehttp://www.worldmapper.org
A world view of some ‘wellbeing’ indicatorswww.worldmapper.org - The world by income. CAVEAT #1 – what is the quality of this data?
http://www.gapminder.org -OECD data
Animated population pyramids (Creator: Martin Ralphs, Statistics NZ)
- using census data (over time and projecting into the future)
Country population structures and changes
Country ‘wellbeing’ indicators(a) Interrogating economic indicators• Looking inside the Consumer’s Price Index:www.destatis.de Federal Statistical Office of Germany
(b) Viewing geographic differences - interactive atlaswww.stat.si/eng/iatlas.asp Statistical Office of the Republic of Slovenia
Displaying multiple national variables Data: Statistics New Zealand Labour Force Synthetic Unit Record File (Labour Force SURF)
Income by hours worked by age by sex (R graphic)
What about at the city level?(a) Mapping population (Census) data - Vienna City Map(b) Commuting data – Home-to-work, Work-to-home • Large and complex 2006 Census tables• By different sized areas in New Zealand:
o Territorial Authority by Territorial Authority = over 5,000 cellso Area Unit by Area Unit = more than 3 million cells!o Meshblock by Meshblock… forget it (2 billion cells).
Far North District
Whangarei District
Kaipara District
Rodney District
North Shore City
Waitakere City
Auckland City
Manukau City
Total Akld Papakura District
Far North District 16860 396 36 30 36 12 108 42 201 9Whangarei District 285 26379 276 75 57 36 171 54 321 12Kaipara District 42 327 5931 315 33 12 69 27 138 0Rodney District 30 48 126 21183 6822 1701 5706 627 14856 54North Shore City 48 63 24 1755 58383 1905 28188 2604 91077 180Waitakere City 48 48 12 1155 4332 31794 30957 3288 70371 258Auckland City 201 141 39 738 7257 6183 140517 16023 169983 942Manukau City 75 69 18 282 1824 1050 40881 66210 109962 3384Total AKld 372 321 93 3930 71793 40932 240543 88122 441396 4764Papakura District 9 15 0 33 177 84 3894 5079 9231 6567Franklin District 12 15 3 48 171 99 3117 3720 7110 1869
Integrating maps and data analysesExciting new software available – GeoVistaCAVEAT #2 – visualisations can be too busy?
Frank Hardisty,
Auckland Ethnicity
Visualisations for others-Business tool box
When not to use graphics
• When distinct numbers are of interest e.g. foreign exchange rates, GDP, etc.
• When tables give a more accurate picture e.g. to convey values of summary statistics
• When you are not sure of the data quality
• To mislead the reader
NOTE: If a visual looks slick and expensive, most people will assume it is correct.
