Each square in the example above is an XY chart constructed by taking x and y values from two table columns. ‘col1′ provides x values or all the charts in its column, ‘col2′ provides x values for all the charts in its column, and so on. ‘col1′ also provides the y values for all the charts in its row
Entrance correlation charts provide a quick, visual way to look for relations between table columns.
It is very easy to see that something is going between col1 and col2 in this (cooked up!) example, because the dots clearly fall on a line. We also see something is going on with col3 and col4, because they seem to have arranged themselves on a curve.
Click on the chart to see the script that made it. Here is the command to draw the chart:
PLOT CORRELATION
SMALL BLUE CIRCLE
WITH
FRAME 30 30 430 430
SELECT
col1, col2, col3, col4 FROM foof;
With one thousand rows in the sample, the results are even more striking:
Notice that correlation charts can reveal relationships that are non-linear. For example, here is a correlation plot for
z = w + x*x + y*y*y:
