17.2.3 Scatter Matrices


A scatter matrix is a pair-wise scatter plot of several variables presented in a matrix format. It can be used to determine whether the variables are correlated and whether the correlation is positive or negative.

Scatter Matrices 01.png


Creating Scatter Matrices

To create a scatter matrix:

  1. Highlight at least two worksheet columns(or a range from at least two columns).
  2. Select Plot: 2Ds: Scatter Matrix from the Origin menu or click the Scatter Matrix buttonButton Scatter Matrix.png on the 2D Graphs toolbar.
  3. The Plotting: plot_matrix dialog opens. Customize the options in this dialog and click OK to create the scatter matrix plot.

Each selected column (or a range of it) is plotted against every other selected column (or a range of it) as a scatter plot layer and all layers are presented in a matrix format in the graph.

Setting in the dialog

You can use its dialog to control the creation of the scatter matrix.

Scatter Matrices dialog.png

Input

Input Specify the input data range. At least select two Y worksheet columns (or a range from at least two Y columns).
Grouping Range

Specify the grouping range. The scatters will be colored by the level in the grouping range.

Matrix Profile

Matrix Display

The matrix format can be one of four arrangements

Square

Matrix Display Square.png

Upper Triangular

Matrix Display Upper Triangular.png

Lower Triangular

Matrix Display Lower Triangular.png

Mixed

When select this option, Additional Statistics Info in will be available. The matrix display is controlled by the position of these additional statistics info.

Additional Statistics Info in When select Mixed in Matrix Display, it will be available. Specify the position of the additional statistics info.
  • Lower Triangular
The additional statistics info shows in lower triangular, the Scatter plots will show in the upper triangular.
  • Lower Triangular
The additional statistics info shows in upper triangular, the Scatter plots will show in the lower triangular.
Show in Diagonal Cells
  • <None>
Box Chart or Histogram will not be displayed in the diagonal cells
  • Box Chart
Display the Box charts in the diagonal cells.
  • Histogram
Display the Histogram plots in the diagonal cells.
  • Histogram with Distribution Curve
Display the Histogram plots with distribution curve in the diagonal cells.
  • Distribution Curve
Display the distribution curve in the diagonal cells.
  • Distribution Curve with Fill
Display the distribution curve filling to bottom in the diagonal cells.
Variables in Diagonal Cells

Specify whether to display the long name of columns from source data.

Show Tick and Label

The way of tick and labels arrangement can be one of the options below

None

Ticks None.png

All

Ticks All.png

Alternate

Ticks Alternate.png

Bottom&Left

Ticks Bottom&Left.png

Bottom&Right

Ticks Bottom&Right.png

Top&Left

Ticks Top&Left.png

Top&Right

Ticks Top&Right.png

Gap (in % of Page Dimension)

Specify whether to show gap between each layer. Type a value in the box to control the spacing between the layers in units of % of the width.

Options

Confidence Ellipse

If this is checked, a confidence ellipse will be drawn for each graph based on the chosen confidence level.

Confidence Level in % This is only available when Confidence Ellipse is checked. Use it to specify the confidence level in percentage for the confidence ellipses. This value must be greater than 0 and less than 100.
Linear Fit

Perform a linear fit to each pair of variables.

  • When this box is checked, the fitted line and the adjusted R^2 value will be added to each scatter graph.
  • Beginning with Origin 2019, Adj. R-square values, plus Pearson's r values if selected, are outputted to a new sheet named as ScatterMatrixStatsN.

Additional Statistics

Pearson's r Add a text label for Pearson's r (correlation coefficient) value, to each scatter plot.
Custom Pearson's r Label Custom the label of Pearson's r info, when Pearson's r option is checked.
Adj. R-Square Add a text label for adjusted R-squared value resulting from the linear fit, to each scatter plot. By default, the option is selected but dimmed and only becomes editable when the Linear Fit option is selected (when dimmed, no text label is created).
Custom Adj. R-Square Label Custom the label of Adj. R-Square info, when Adj. R-Square option is checked.
Significant Mark Specify whether show the significant mark for additional statistics info
Significant Level It's default value are 0.05 0.01 0.001. It means there is 3 levels, if correlation coefficient<0.001, significant mark is ***, 0.001<correlation coefficient<0.01, significant mark is **, and so on, when value >0.05, mark is empty
Label Size by Value This option is available when
  • Not set Grouping Range for the input dataset.
  • Pearson's r and Adj. R-Square option, one of them is selected, but not selected both.

In such case, it will vary the size of label according to value.

Exclude Missing Values Listwise

Specify whether to exclude missing values (listwise). That is, exclude the entire row for all datasets if there are any missing values in this row.

Graph Templates

Scatter Specify the Template for Scatter plot output.
Box Specify the Template for Box charts output. This option would be deactivated when None is selected in Show in Diagonal Cells under Matrix Profile node.
Histogram Specify the Template for Histogram plot output. This option would be deactivated when None is selected in Show in Diagonal Cells under Matrix Profile node.

Output Results

This determines where the calculated data for the ellipse and the fit are stored. The default location is a new worksheet (<new>) within the source workbook ([<input>]).

Note: selecting N datasets or ranges will result in N^2-N graphs and an even larger number of datasets. Selecting a large number of datasets or ranges may increase the computation time dramatically and produce small graphs.

From Origin 2023b, if the Grouping Range is specified, it will output the result (including Confidence Ellipse, Linear Fit, Statistics info) for each individual group. In the diagonal cells, the color of plots is also following to the group color of the Scatter plots.