The "Dashboard of Sustainability" is a free, non-commercial software (download the software and the detailed 14-page printable manual) that allows to present complex relationships between economic, social and environmental issues in a highly communicative format.

A new function added in August 2005 is variable weighting based on performance: Indicators for which a given country scores very badly are given a higher weight. Test this function by pressing "V" (like "variable"; press V again to go back to normal). The degree of variation can be adjusted by pressing + and - on the NumPad. Click here for an example.

Finding "surprises" in the indicator set:
Both the "From Rio to Jo'burg" and the "Millennium Development Goals" versions are complex enough to allow for an in-depth analysis based on known correlations between the indicators. The respective function, called "SURPRISES", can be reached via the Export menu by clicking on "Outliers" (see right picture). The following dialog will appear:

exp_su_d.gif (4282 bytes)

After clicking "Start", the Dashboard will check for each country whether the indicators match expected patterns.

 exp_surp.gif (13729 bytes)
How does it work? As the linkage analysis shows (icon ic_linka.gif (1040 bytes)), practically all indicators are more or less strongly correlated with the others. It is therefore possible to estimate a missing value for any given indicator on the basis of available indicators and known correlations. The Dashboard takes the average of all expected values weighted by the absolute value of the correlation coefficient. Especially in embedded mode, i.e. using all countries, the result of this calculation deviates typically only +- 100 points on the 0...1000 scale. You can test this feature by holding Shift and right-clicking into an indicator legend (see right picture): exp_surp.gif (13729 bytes)

What the "Surprises" function does is to compare systematically the real (statistically collected) indicator value with the expected one, and to flag those that show a large difference. Here is an example taken from the MDG set:

Real Expect RegrQu DIFF Indicator
points for this indicator based on collected data (0=worst, 1000=best country) expected value calculated on the basis of known correlations between all indicators average quality of regression coefficients * 1000 (e.g. r=-0.535 ->535 quality points) difference between observed and expected value; negative: indicator value is surprisingly bad  
Bangladesh        
- positive:        
997 332 204 665 Energy efficiency
960 576 477 384 Enrollment
- negative:        
95 496 525 -401 Attended births
352 666 490 -314 Youth literacy
mdg_fvb.gif (7787 bytes) The result for "Youth literacy" should be interpreted as follows: Bangladesh scored 352 points, which is not bad for a Developing Country but 314 points worse than we expected based on the correlations to other indicators.

For the overall index "Millennium Development Goals", Bangladesh is in the middle range with 505 points, but the Youth literacy rate is far below the regression line. You can check this more systematically, i.e. for all indicators, with the Linkage Analysis, and it becomes obvious that Bangladesh's Youth Literacy Rate is most of the time far below the regression line. This observation is a "surprise", and needs further interpretation. It may be an error in the data, but it may also be a result of a specific problem that could eventually be solved knowing the cause. Strangely enough, the value for School enrollment is 384 better than expected...!

Click here to see the full results for the Millennium Development Goals set. Note that this function is relatively slow, but you can interrupt it by pressing Shift of Escape or the right mouse button. The results will be displayed by Notepad, although it is recommended to paste them into a spreadsheet for better viewing.