22.214.171.124 The (Plot Details) Parallel Tab
The (Plot Details) Parallel tab provides controls for styling parallel plots. This includes controls for color transparency and for "curvature" (the smoothing effect commonly seen in what are termed "alluvial diagrams").
Select a line style from this drop-down list. Note that when plotting large datasets (ie. when lines overlay one another), changing this setting may have no visible effect.
- You can customize the dash patterns using the Origin Dash Lines group controls on the Graph tab of the Options dialog box (Tools: Options).
- At low screen resolutions or in small windows, the dashed lines may appear solid. However, a printout should draw the line correctly.
Type or select the desired line width in this combination box. The line width is measured in points, where 1 point=1/72 inch.
Select your color using Origin's Color Chooser. Note that, depending on plot type, there are up to three tabs -- "Single", "By Points" or "By Plots" -- on the Color Chooser.
Typically, when creating a parallel plot, you will map color using a column of numeric values, by clicking the By Points tab and choosing the color column from the Map drop-down.
When creating a parallel sets plot, you will apply color using a column of categorical values. For instance, you might assign plot color by gender (e.g. "male"=blue, "female"=pink) by clicking the By Points tab and from the Index drop-down, point to the worksheet column containing gender data.
For more information, see Using a Dataset to Control Plot Color.
When the Color is set to Map, line colors for each value can be further customized in the Colormap tab, while Color is set to Increment or Index, pipe colors for each category can be further defined in the Color List tab.
Move the slider, or type a desired integer from 0 to 100 in the combination box. Note that 0 means no transparency, 100 is fully transparent.
Note that you can use the Apply button to view slider effects and not have to close the Plot Details dialog.
Move the slider, or enter an integer integer between 0 and 100 in the combination box. Note that 0 means no curvature, 100 is maximum curvature.
This check box on the Parallel tab of Plot Details, modifies the drawing of the graph in significant ways.
- When Combine Sets is disabled (default), the categories and color-split of the left-most axis are carried through to the right-most axis.
- When Combine Sets is enabled, you are creating sub-plots defined by each pair of vertical axes.
Which setting you choose will depend on what you want to emphasize. In the above example, a dataset of high-school students is studied for the effect of community-type and gender on their goals as students. When Combine Sets is disabled, you can more clearly see the effect of community-type (Urban=yellow, Suburban=blue, Rural=red) on Goals, as the colors assigned to these categories carry through the entire graph.
Conversely, when Combine Sets is enabled, it is more difficult to pull out the effect of community-type on Goals but the effect of Gender on Goals becomes clearer. Thus, it makes sense to toggle this setting on and off to see how it shifts emphasis.
There are a couple of other things that you can do to help you to see trends in Parallel Plots. One is to click on a plot in the graph page. This will fade all other plots in the page. Another is to use the check boxes in the Object Manager to temporarily hide or show plots in the graph page.
Plotting Order Within Categories
When plotting numeric data, plotting order of data points is by row index (row number). When plotting categorical data, plotting order within a category is determined by the sort order in the categorical column. If data are unsorted, plot order is determined by order of appearance within the column.
When you enable Combine Sets, plotting order is determined only by the sort order in adjacent columns. Note that plotting order is from the base of the graph upward.
Gap Between Categories
Controls space between the categories on each vertical axis. Experimenting with this setting may help to create better visual separation of categories. Not applicable to variables plotted on a continuous scale.