Look-and-Feel

Data Desk is a desktop environment. If you guess that objects on the desktop work the way they do on other desktops, you’ll probably be right. Go ahead and experiment. You can’t really do much damage.

Much of the look-and-feel of Data Desk can be controled with the preferences found in the {Data Desk}Preferences menu.

Working with Data Desk

To perform most data analysis or display operations, select a command for the plot or analysis from the Plot or Calc menu. Data Desk will open a blank window for that result and invite you to drag variables into it. For windows that do not differentiate the role of a variable (e.g. histograms, pie charts, summary statisitics), just drag a variable into the window and drop it there. For windows that differentiate roles, drag variables to the indicated locations within the window. Thus, for example, place a variable you wish to plot on the y-axis of a scatterplot on that axis and a variable you wish to plot on the x-axis on that axis. Similarly, drag the response variable in a regression to the designated part of the regression table and as many predictor variables as you wish into the designated area for predictors.

Windows update immedately when you drop variables.

You can also modify an existing plot or table by dragging variable over the names of variables currently there to replace them. In some windows, you can add additional variables to the display (e.g. rotating plot, parallel coordinate plots, boxplots side-by-side) or analysis (e.g. regression, linear models).

Interface Scale

Adjust the interface scale to suit your way of working. The smallest interface leads to small objects, and thus less crowding on the desktop. The largest scales are easier to see and read.

Frame Rate

Animations, such as Rotating Plots work by showing a sequence of pictures to create the animation effect. Here, you control the frame rate, from slow to far faster than you can see.

Color Scheme

Traditional Data Desk plots display white or colored symbols and lines against a black background. This color scheme is particularly effective for showing details in plots. But it is not what most folks are used to seeing

The Dark on Light color scheme is the standard one on Data Desk 7.

Desktop Pattern

Choose the background pattern you like for the Data Desk desktop

Numbers to Text

Data Desk offers full control over how numbers are displayed and worked with.

Keyboard

Specify the keyboard you are using.

Categories Limit

Commands that recognize categories in variables have a safety limit to the number of categories just in case you accidentally use a quantitative variable rather than a categorical variable. Set that limit here.

Drag-and-Drop

Most plots and statistical analyses are specified by first selecting the commands from the Plot or Calc menus. Data Desk opens an appropriate type of window, which invites you to drag the icons of variables into parts of the window. Thus, for exmple, you can drag variables onto the y-axis and x-axis of a scatterplot.

Drag-and-Drop also works to modify existing plots and analyses. Drag variables to add them to analyses or to replace existing variables in plots or analyses.

HyperViews

Data Desk’s windows offer HyperView menus that offer modifications to the window and suggest related plots or analyses. HyperView menus have a built-in knowledge of how statistics methods work together so they can place related analysis steps at your fingertips. A window’s HyperView menu pops up when you press the arrow at the left side of the window’s title bar. Other HyperView menus are attached to parts of the window. When the mouse cursor changes to a pointer hand, it is over a HyperView menu; press the mouse button to pop up the menu.

Most Data Desk results windows suggest additional or alternative analyses or plots. These might be checks on the underlying assumptions of a procedure (such as a histogram to check how a variable is distributed) or they might be naturally related analyses (a frequency breakdown to provide the counts and percentages graphed in a pie chart). For example, when you press the mouse button over an axis label in a scatterplot, the HyperView menu that pops up offers to locate the icon, make a histogram, or make a normal probability plot of the variable plotted on the axis. If you press the mouse button over a correlation coefficient in a correlation table, the HyperView menu suggests a scatterplot of the underlying variables.

Global HyperView Menus

Global HyperView menus are attached to the window as a whole. Most Data Desk windows have a submenu arrow located in the upper left corner of the title bar next to the close box. The HyperView menu attached to that arrow suggests general actions related to the analysis or display in the window.

Context-Sensitive HyperView menus

Context-Sensitive HyperView menus are attached to specific parts of the plot or table and suggest analyses or plots related to those parts.

These HyperView menus can be more context-specific than global HyperView menus. In plots, for example, the HyperView menu attached to the axis labels usually offer to locate the icon of the displayed variable or to show it in a simple one-variable display, such as a histogram. In most tables, the HyperView menu attached to the test statistic usually offers a display to check if the assumptions are valid.

Variable-ID First Command Specification

A more direct way to specify commands to Data Desk is to select variables first and then choose a command. The first variable you click on will be selected as a Y variable (and will then be the variable described or modelled). Hold the shift key and click on more variables to select them as X variables (which will be used as predictors or to graph on the horizontal axes of plots.) Hold the Option key down to select more variables as Y’s (useful for multivariate methods).

Plot Links

Data Desk places each plot in its own window so you can see multiple views of your data. Plots are linked together. When you select points or parts of a plot, the corresponding cases highlight immediately in all displays.

Window Updating

Most results windows (plots and tables) offer to update whenever any of their arguments (variables or values) changes. Windows indicate that they no longer correspond to their underlying arguments by changing the Global HyperView in the upper left corner to a red exclamation mark. Click on the mark to either update the window in place or make a new, updated copy, leaving the original as it is.

Some operations update windows immediately. Case-level changes’selection, assigning a color, or assigning a display symbol’change immediately. Thus, the Brush, and Knife tools create a selection animation, changing selection in all windows as they are moved. These aspects of displys’seletion, color, and symbol’are always kept consistent across all displays of data in the same relation.

You can Turn on Automatic Update in a windows global HyperView. When in that state, the window’s HyperView menu appears gray. All changes that would lead to setting the red exclamation mark will instead update immediately. This setting can be used to create animations, for example, by using a Slider in a derived variable, and then using that derived variable in a plot or analysis. If sliding the control changes the value of the derived variable, all plots and analyses that depend on it will update immediately.

Sliders

For re-expression

For variable mixing

Results Folder

Each plot or analysis generates one or more results. The Results folder (ordinarily in the file cabinet at the upper right of the desktop) holds these icons. Data Desk saves all of your results (until you discard them in the Trash), so you can always return to a previous point in your analysis and start a new analysis path from there.

All Data Desk icons have associated history information. The Data>Info command displays that information. The information inludes the date and time of creation and most recent modification. The Results folder is kept in order of creation date and time.

Preserve Results History as a ‘Lab Notebook’

The icons in the Results folder hold displays and analyses. However, Data Desk will offer to update those results whenever their arguments change, and you may choose to make some updates or changes. For example, you might drag an additional predictor into a regression table. Thus, the Results folder is not, in its basic setting, a perfect record of your analysis path. If you would like to preserve such a record choose the Preserve Results History option in the Preferences. With this setting, any time a result is changed, Data Desk will keep the original result and make a copy of it as a new result (with a new icon and creation date) in the Results folder. The Results folder will then be a faithful record of everything you did during your analysis.

This setting may have some unexpected consequences. For example, suppose you have performed a regression and plotted residuals. Now when you drag a new predictor into your model, Data Desk will make a fresh copy of the regression table. You will now have two regression tables. The residual plot, which is a child of the original analysis, will not offer to update because its ‘parent’ has not been updated. This may or may not be what you intended or expected.

Full Screen Mode

The Data Desk>Enter Full Screen command expands the Data Desk desktop to cover your full screen and makes the Data Desk menu bar the only visible menu bar at the top of the screen. Choose Data Desk>Exit Full Screen to restore the previous desktop size.