Togaware DATA MINING
Desktop Survival Guide
by Graham Williams
Google

Brushing

From a data mining perspective we are usually most interested in the relationship between the values of the variables for observations for which it does or does not rain tomorrow (using the variable RainTomrrow). We can have these two different values highlighted in different colours very easily. From the Tools menu we can choose Automatic Brushing to have the window shown in Figure 7.7 displayed.

Figure 7.7: GGobi's automatic brushing.
Image explore:ggobi_autobrush

From the variables list at the top of the resulting window choose RainTomorrow after scrolling down through the list. Notice that the number ranges in the lower colour map change to reflect the range of values associated with the chosen variable. For RainTomorrow which has only the values 0 and 1, any observations having RainTomorrow as $0$ will be coloured purple, whilst those having it as $1$ will be coloured yellow.

We click on the Apply button for the automatic brushing to take effect. Any plots that GGobi is currently displaying (and any new plots we cause to be displayed) will colour the observations appropriately, as in Figure 7.8. This colouring of points, across multiple plots, is referred to as brushing.

Figure 7.8: Automatic brushing of multiple scatterplots using GGobi.
Image explore:ggobi_scatter_evsun_touched_br       Image explore:ggobi_scatter_maxmintemp_notcurrent_touched_br

Our plots can be somewhat more colourful by choosing a numeric variable, like Sunshine, as the choice for automatic brushing. We can see the effect in Figure 7.9.

Figure 7.9: Colourful brushing of multiple scatterplots.
Image explore:ggobi_scatter_evsun_touched_br_sun       Image explore:ggobi_scatter_maxmintemp_notcurrent_touched_br_sun

Copyright © Togaware Pty Ltd
Support further development through the purchase of the PDF version of the book.
The PDF version is a formatted comprehensive draft book (with over 800 pages).
Brought to you by Togaware. This page generated: Sunday, 22 August 2010