The Explorer User Interface

The Explorer

The SAS Visual Analytics Explorer user interface
1 The application bar enables you to return to the home page and to access other parts of SAS Visual Analytics and other SAS applications that integrate with the home page. You can access your recently created or viewed reports, explorations, stored processes, data queries, or other objects in your recent history. Buttons are displayed for each open application.
2 The menu bar offers common tasks, such as creating a new exploration.
3 The toolbar enables you to manage your explorations and visualizations.
4 The Data pane enables you to manage the data that is used in your visualizations.
5 The data properties table enables you to set data item properties.
6 The workspace displays one or more visualizations.
7 The right pane’s tabs enable you to set properties and data roles, create filters and ranks, set global parameter values, and use comments.
8 The dock contains any minimized visualizations.

Menus and Toolbars

From the SAS Visual Statistics main menu, you are able to access all of the features of the application.
The SAS Visual Statistics toolbar enables you to quickly access frequently used tasks:
Icon
Description
create a new visual exploration
Creates a new exploration.
maximize model workspace
Maximizes the modeling workspace.
return workspace to the default
Returns the modeling workspace to the default view.
linear regression
Creates a linear regression model.
logistic regression
Creates a logistic regression model.
generalized linear
Creates a generalized linear model (GLM).
decision tree
Creates a decision tree model.
new cluster
Creates a cluster model.
compare models
Compares two or more models.

The Data Pane

The Data pane enables you to access all of the variables in your data set. Variables are sorted into Category and Measure variable groups. By default, character variables are considered category variables and numeric variables are considered measure variables. Numeric variables with a finite number of levels can be converted into category variables.
You can create variable interactions and calculated items in the Data pane.
When you enter something into the Search data field, only those variables that contain the search term are displayed. Search terms are not case sensitive.
To collapse the variable groups, click collapse all to the left of the Search data field. To expand the variable groups, click expand all to the left of the Search data field.
The Data pane drop-down list icon, Data Pane Drop-Down List, is located in the upper right corner of the Data pane. The following items are of interest, but not a complete list of options:
  • New Interaction Effect enables you to create interaction effects to use in your visualizations.
    New Interaction Effect window
    In the New Interaction Effect window, you move variables from the Available columns area to the Effect elements area. You can drag and drop a variable, double-click a variable, or use the arrows in the center of the window. After you have moved variables to the Effect elements area, click Create to create a single interaction. Alternatively, click Menu Arrow Down icon to specify whether you want to Create one interaction or Create two-way interactions.
    When you choose Create two-way interactions, all of the possible pairs of interactions for the selected variables are created, except for square terms. To create a square term, move two copies of a variable into the Effect elements area, and click Create.
  • Data Properties displays the name, classification, data type, model type, and format for each variable in the data set.
  • Measure Details opens the Measure Details window. The Measure Details window provides summary statistics and a histogram for each of the measure variables.
  • Show/Hides Items enables you to specify what variables are displayed in the Data pane. Variables in the Visible items area are displayed in the Data pane. Variables in the Hidden items area are not displayed.
    To move variables from one area to the other, drag and drop the variable, double-click a variable, or use the arrows in the center of the window. You can move multiple variables by selecting them first, and then either dragging and dropping them or using the arrows.
    Click OK to close the Show or Hide Items window and save your changes.
  • Sort Items enables you to specify whether you want to sort the variables in ascending or descending order.

The Right Pane

The right pane contains the Roles, Properties, and Filters tabs. Together, these three tabs define the modeling parameters. The Roles tab specifies what variables are used in the model and their purposes in the model. The Properties tab enables you to specify features that are unique to each model. The Filters tab enables you to subset the data that is modeled.
Tab
Description
Data Items
Roles
Enables you to add variables to your model.
From the Data pane, drag and drop a variable that you want to use to a role on the Roles tab. Alternatively, you can select several variables and drag and drop them onto the model pane. In this case, each variable is assigned to the first valid and available role.
If there is no response variable, then the first valid variable is assigned to the Response role. This method never assigns Group By, Frequency, Filter, or Weight variables. You can use the Menu Arrow Down icon to add or remove variables from each individual field.
Filter
Filters
Enables you to specify variables that are used to filter the data set. You can filter on category variables, measure variables, or both. To add a filter variable, drag and drop that variable from the Data pane to the Filters tab or use the Drop Down Menu icon. To remove a filter variable, click Close next to that variable’s name.
Ranking
Ranks
Enables you to specify rank variables. Rank variables subset the data in your visualization to display either the top or the bottom aggregated value for a category. To add a rank variable, drag and drop that variable from the Data pane onto the Ranks tab.
Properties
Properties
Enables you to specify features of the model. The available options vary based on the selected model.
Comments
Comments
Enables you to enter comments about the current model.
Parameters
Parameters
Enables you to view any parameters that affect a calculated item that is used by the current model.

The Model Pane

The model pane contains the modeling results and plots. Because the windows available depend on the selected model, this section focuses on the common elements for all models. Specific information for each model is available in the SAS Visual Analytics: User’s Guide.
The summary bar displays the response variable, model evaluation criterion (when available), and the number of used and unused observations in the model. To see all of the available model evaluation criteria, click the name of the current model evaluation criterion in the summary bar to open a pop-up menu.
Summary Bar
On the right side of the summary bar, there is the Details View Icon icon. Click Details View Icon to open the details table at the bottom of the model pane.
An example details table from a decision tree model is shown below. The specific information in each details table varies based on the model.
Details Table
To maximize a window, click Maximize Button. This hides all of the other windows in the model pane, but does not hide the details table. To restore the default view, click Restore View.
The Gallery contains a sample screenshot and description for each window.
Window Name
Available In
Fit Summary
Linear, Logistic, GLM
Residual Plot
Linear, Logistic, GLM
Assessment
Linear, Logistic, GLM, Decision Tree
Influence Plot
Linear and Logistic
Tree
Decision Tree
Leaf Statistics
Decision Tree
Cluster Matrix
Cluster
Parallel Coordinates
Cluster
In every window, if you position your mouse pointer over an object, a tooltip provides specific information about that object. The information varies based on the plot that is displayed. For example, in the following image, the tooltip shows the percentile (in the bin), observed average value in that bin, and the number of observations in that bin.
Assessment Window with Tooltip
Whenever a range of values is shown or selected, the interval is half-open. The minimum value is included in the interval. The maximum value is excluded from the interval. This affects heat maps, the parallel coordinates plot, and any other displayed or selected interval.