Graph: Layout (Matrix Plots)

From the Help - Animated Overviews menu, select How To: Dynamic "Data-Graphs" Links to view a multimedia presentation of the use of dynamic features to automatically update graphs.

In the Graph Options dialog tree view, select Layout (under Graph) to access the options described here. Use these options to modify the graph type of the existing Matrix plot. Several options are available for changing the display format of the graph, changing the order of the variables, and treatment of the missing data.

Graph Type. Select the type of matrix plot from the Graph Type list. You can choose between two types of matrix plot layouts:

Square Scatter Matrix. Select Square Scatter Matrix to produce a square scatter matrix plot, plotting all variables against each other. Note that, in order to include all variables (see Variables order) in the plot, you must specify as many Rows (see below) as you have variables; otherwise a truncated list of variables will be shown in the matrix plot.

Rect. Scatter Matrix. Select Rect. Scatter Matrix to produce a rectangular scatter matrix plot, plotting n Rows variables against m Columns variables (see Variables order below), producing an n by m matrix of plots; specify the number of Rows and Columns in the respective fields below.

Marker point reduction. Use the Marker point reduction option to specify an optimization for rendering graphs with large amounts of data. Usually, this option does not affect the overall appearance of the graph, it simply allows for quicker display of the graph when large (i.e., over 100,000 cases) data sets are involved. The default value is Standard, and for most data sets this is the desired setting. With this setting, every point is plotted. When two cases have the same coordinates, the second case is simply plotted over the next case. When Fast is selected, multiple points with the same coordinates are skipped (i.e., only one point with that set of coordinates would be plotted; although all cases are shown in the graph data editor). This considerably reduces the graph drawing time. The Fast option is ideal for larger data sets. The Aggressive option allows for two markers to be considered overlapping even if there is a slight (one pixel) difference between them in the graph. This option still provides a good representation of the data while drawing the graph about four times quicker.

Visual differences might exist when point labels are used in the graph or when the markers are different (e.g., case state markers are used).  For example, when Standard is used in conjunction with point labels (i.e., text values), the labels of overlapping points cannot be read. When Fast is used in conjunction with point labels, the point labels will be easier to read, because only one of the cases (the last one) with those coordinates is actually plotted. Hence, there is only one point label for that location. Another visual difference occurs when point markers are different for subsets of cases (i.e., case states markers are used). When the Standard option is used and points with different markers overlap, you would expect to see fragments of the markers from previous cases (e.g., fragments of a square are still visible when overprinted by a diamond). With the Fast option, only the marker associated with the last case to have those coordinates will be shown.

Layout. Use the Layout group to modify the layout of the matrix plot and the manner in which you want to treat the missing data when the matrix plot is recomputed.

Rows/Columns. Use the Rows/Columns fields to specify the n number of Rows (for the Square Scatter Matrix and Rect. Scatter Matrix) and m number of Columns (for the Rect. Scatter Matrix) for the n by m matrix plot. Note that for the Square Scatter Matrix only a single dimension (Rows) n can be specified for the n by n plot. If the total number of rows and columns is less than the number of variables (see Variables order) then not all variables will be shown in the plot.

Missing Data Deletion. Use the Missing Data Deletion group box to choose between Pairwise and Casewise deletion of missing data. When the Pairwise option button is selected, cases will be excluded only from any calculations involving variables for which they have missing data. When the Casewise option button is selected, then only cases that do not contain any missing data for any of the variables selected for the graph will be included in the matrix plot.

Variables order. The Variables order box displays all the variables that are used in the current matrix plot. To change the order of variables, first select the respective variable and "drag"  the variable(s) up or down to change the order of the variables to be used in the modified matrix plot.

Original. Click the Original button to restore the order in which the variables were originally used in the matrix plot.

Data update. Use the options under Data update to specify the update method applied to the graph when data change in the data file.

Automatic. Select this option button to specify that the graph data is to be updated automatically when changes are made to the data source. This will be the default setting.

Manual. Select this option button to specify that the graph data is to be updated  manually when the user explicitly requests an update of the graph data.

Locked. Select this option button to specify that the graph data is not to be updated.

Details. Click the Details button to display the Data Update Details dialog, which contains more detailed options to adjust the spreadsheet data source and variable selection used for the graph.

Don't ask before performing automatic updates on modified graph. Select this check box to specify that you don't want to be warned before the automatic update of the graph takes place.

Update on OK. If the Data update mode is set to Manual and this check box is selected, a prompt will be displayed when the source data change, and the graph will be updated when you click the OK button.

Reconnect on load. Select this option to control how the graph reloads the data from the spreadsheet. The options are Always, If spreadsheet open, Always ask, and Never.

Update spreadsheet case states. When this check box is selected, the changes of the case states (including the case labels, case excluded, case hidden, Marker color, etc.) made in the graph will be automatically updated to the spreadsheet, so that the spreadsheet that the current graph is based upon will have the same case states as in the graph.

Update Spreadsheet on OK. If this check box is selected, the current graph case states will be updated in the connected spreadsheet when the dialog is closed. If the Update Spreadsheet Case States check box is cleared, then this option can be used to perform a manual update of the spreadsheet case states.

Moveable point labels. Click this button to display the Moveable Labels dialog, where you can set options for moveable labels.