Parallel coordinates are a way to plot multivariate data. They are primarily used to visualize the relationships between many numeric variables. The chart gives each variable its own axis and positions it in parallel to the other axes, with the data points plotted as a series of lines along each axis.
Data format
The variable order of the parallel coordinates chart will determine how the axes are arranged. This is important to note because it is much easier to detect patterns between variables that are positioned adjacent to each other.
The axes are ordered exactly as the variables are ordered in the variables list. Re-ordering the axes in a meaningful way will make it easier to discover patterns and correlations.
Interpretation
A parallel coordinates chart can appear messy and cluttered. Here are some things to look for when interpreting the results.
-
Patterns between colored lines Setting the color variable (Data > Data Source > Color variable in the object inspector
) will color-code each line according to the assigned variable. This allows you to more easily detect patterns and correlations.
- Clusters on a single axis Data points may be clustered around certain areas of an axis, which may be particularly noteworthy if the clusters are also differentiated by color on the chart.
- Patterns between axes Correlations between axes adjacent to each other indicate that the variables are in some way connected.
Object Inspector Options
The following is an explanation of the options available in the Object Inspector for this specific visualization. Refer to Visualization Options for general chart formatting options.
Chart
APPEARANCE
- Variables Variables to show in the chart. These must come from the same data set. Each variable will be shown on a separate axis.
- Color variable Optional variable used to color the lines in the chart. It must be a variable from the same data set as Variables, and can even be one of the variables included in Variables. If no variable is supplied, then all the lines in the chart will shown in the same color.
- Variable names Use variable names instead of variable labels to label the axes of the chart.
- Tidy labels Extract common prefixes from axes labels.
- Reverse axes Show smallest values at the top of the y-axis.
- Enable interactions When interactions are enabled users can click and drag on chart to select lines. However, for large data sets it may be more efficient to turn off interactions.
- Render lines progressively Lines are gradually added to the chart, so the user can see when multiple overlapping lines are added.
- Line opacity Set transparency of lines to a numeric value between 0 (clear) to 1 (opaque). Note that overlapping lines of multiple colors will always be shown as a dark line (black hue).
- Line colors Select a color palette to use if Color variable is provided, or a single color if not.
- Rotate variable label Show variable labels at perpendicularly. This is useful for displaying longer labels. You may also need to adjust margin dimensions.
Acknowledgments
Uses the parcoords package by timelyportfolio.