Correlation Plots
Correlation Plots are plots used to visualize correlations between ensemble parameters and summary result vectors. Several visualizations can be created, including tornado plots, correlation matrices, and cross plots.
Creating a Correlation Plot
The most convenient way to create a Correlation Plot at a given time step is often to right-click an Ensemble Plot and invoke the submenu shown below.
Alternatively, Correlation Plots can be created by right-clicking the Correlation Plots item in Plot Project Tree.
Cross Plot Filter
The realizations are displayed in a cross plot with result parameter vs summary value. A filter on x and y axis can be set. The realization matching the filter can be used to create a new ensemble. This option is available on the right-click menu of a correlation plot.
A new ensemble can be created based on the filtered cases in the right-click menu of the cross plot. A histogram of the ensemble parameter can also be created.
Selecting Parameters in a Correlation Report
In a correlation report, click a bar in the tornado plot to select that ensemble parameter. The other plots in the report are updated: the parameter is highlighted in the correlation matrix, and the cross plot shows the selected parameter against the summary vector. Clicking a cell in the correlation matrix selects the parameter and summary vector in the same way.
History Values
When the selected summary vector has a matching history vector in the ensemble (e.g. WOPRH for WOPR), the history data is shown together with the simulated results:
- Cross plot – A dashed horizontal line marks the history value at the selected time step. The y-axis range is extended to include the line.
- Correlation report – The history curve is added to the summary plot, and is updated when another summary vector is selected in the report. Hovering over the summary plot shows dashed lines with the time and value at the cursor position.
If a history vector is given as summary vector when creating a correlation report, the corresponding simulated vector is used instead.
Specifying Setup and Settings
The Correlation Plots property editor allows the user to specify an extensive number of settings dependent on the actual type of correlation plot. The complete set of settings are listed below according to the individual sections of the property editor.
Correlation Factor Settings
Correlation Plots use the Pearson Correlation Coefficient. This section of the property editor allows for selecting ensemble parameter(s) of choice. Being irrelevant, please note that the section will not display for Correlation Cross Plots and that the field Sort Matrix by Values will not display for tornado plots.
Summary Vector Settings
This section of the property editor allows for setting summary vector(s) and time step. To ease the selection of time step, the Time Step Filter may be used to skip time steps by years, quarters, months, weeks, or days.
Plot Settings
This section of the property editor allows for setting plot title, font sizes, and legend if relevant. Please note that the Colors setting for specifying legend is irrelevant to tornado plots and cross plots and will not display for these types of Correlation Plots.
Cross Plot Parameter Setting
This section of the property editor will only display for cross plots to allow specification of ensemble parameter to cross plot with summary vector(s).
