Scatter Plot

Previous  Terms  Next

TinkerPlots_logoV2-rgb


Scatter Plot

A scatter plot allows you to explore the relationship between two numeric attributes with one displayed on the horizontal axis and the other on the vertical axis.

TinkerPlotsHelp-1-068-082

To make a scatter plot,

1.

Put a numeric attribute on the horizontal axis. From the data cards, drag the name of a numeric attribute onto the horizontal axis of the plot (the lower part of the plot will highlight to indicate when you can drop).

2.

Fully separate the attribute. In the plot, drag any case icon all the way to the right (until there are no bin lines). This will fully separate that attribute horizontally.

3.

Put a numeric attribute on the vertical axis. From the data cards, drag the name of a numeric attribute onto the vertical axis of the plot.

4.

Fully separate the attribute. In the plot, drag any case icon all the way up to fully separate it vertically.

There are various ways in TinkerPlots to produce a line or curve of fit on a scatter plot. You can use the Drawing tool on the lower plot toolbar to draw a line (or curve) yourself. You can also change the scatter plot into a binned scatter plot and use averages to help see a trend. You can also use the Meter tool to create a smoothed line of fit. Several of these methods are demonstrated in the movie "Exploring Relationships 2."

Often with scatter plots, two or more cases will lay exactly on top of one another because they have the same values for both attributes. There is currently no option for "jittering" cases so that you can see them or for creating special plotting symbols that stand for different numbers of cases. There are a few things you can do, however.

To locate overlapping cases, you can pull an icon up and move it temporarily out of the way to see if there are any hidden cases. You can also create bins for one of the attributes. Finally, you can change the icon type to Fuse Rectangular, which will put cases that are close together or the same value into rectangles in which you can see each individual case.


TinkerPlots Help

© 2012 Clifford Konold and Craig D. Miller