Prism can create graphs in the style of classic column graphs directly from a multiple variables data table, without restructuring your data into a Column or Grouped table first. These graphs plot one categorical variable against one continuous variable, and they support the same familiar graph elements you would use on a classic column graph: bars, box and whiskers, violins, error bars, and summary symbols and lines.

This is useful when your data are already in a "tidy" or database-style layout, with one row per observation and a column identifying the group each observation belongs to. That is the format most instruments and analysis pipelines export.

What your data need to look like

You need at least two variables in the same multiple variables data table:

•One categorical variable, which defines the groups. Each unique value in this variable is a level, and each level gets its own position along the categorical axis.

•One continuous variable, which supplies the values that are plotted and summarized.

Prism assigns variable types automatically when it can. If a variable is not being treated the way you expect, set its type manually on the Variable Types tab of the Format Data Table dialog.

Creating the graph

From a multiple variables data table, click New Graph... in the Navigator. The Create New Graph dialog shows a thumbnail for each available graph type. The categorical graph types are:

•Strip Plot: one symbol per row of data, plotted within each level. The multiple variables equivalent of a classic scatter (column) plot.

•Summary Symbol Plot: one symbol per level, positioned at a summary value, with error bars.

•Bar and Strip Plot: bars with the individual data points plotted on top.

•Before-After Bar and Strip Plot: bars, individual points, and connecting lines drawn across levels.

•Before-After Strip Plot: individual points with connecting lines drawn across levels.

•Box and Whisker Chart: a box spanning the interquartile range for each level, with whiskers and a median line.

•Violin Plot: a mirrored kernel density curve for each level, showing the shape of the distribution.

•Bar Chart: one bar per level, extending from a baseline to a summary value, with error bars.

Choose a thumbnail, then confirm the variable assignments. Prism proposes assignments based on the variables in your table: if the first variable is categorical it is placed on the X axis and the first continuous variable on the Y axis, otherwise the first categorical variable is used for the Y axis. You can change these at any time on the Axis Variables tab of the Format Graph dialog.

Placing the categorical variable on the X axis produces a vertical graph. Placing it on the Y axis produces a horizontal graph.

Changing an existing graph

To convert a graph you have already made, use Change Graph Type. The same categorical thumbnails are offered there.

Customizing the graph

Each graph type is simply a starting combination of data objects (GDOs). Every data object can be turned on or off and formatted independently, so you are not limited to the preset combinations. For example, you can enable both violins and box and whiskers on the same graph to produce a combined plot, and control the relative width of each.

•To add, remove, or format the elements on the graph, use the Data Objects section of the Format Graph dialog, or the Graph Inspector.

•To change the order in which the levels appear along the axis, use the Level Order tab.

•To color the graph by a variable rather than a single color, use the color settings for each data object. Color can be defined by a single color, a continuous variable (producing a gradient), or a categorical variable.

If your license does not include categorical graphs

Assigning a categorical variable to an axis requires a Prism Pro or Enterprise license. With other licenses:

•You can still open, view, copy, and export files containing these graphs, so graphs shared by colleagues remain readable.

•You cannot create, modify, or format them. The categorical graph thumbnails and the affected controls are shown but disabled, with an Upgrade indicator and a link to more information.

•If the input data are changed, the graph contents are cleared, because Prism cannot redraw the graph under your current license. A floating note on the graph explains this.

•Saving the file again preserves the graph settings, so the graph is restored when the file is opened with a license that includes the feature.

Compatibility with other versions and formats

•Files containing these graphs cannot be saved in the legacy .pzf or .pzfx formats. Save in .prism format instead. See compatibility issues.

•Earlier versions of Prism, including Prism 11.0 and all Prism 10 versions, do not recognize a categorical axis variable. They open the file with an alert, show the graph as empty, and display a floating note suggesting an update to the latest version of Prism.