Skip to contents

4. Plot Results

Source: vignettes/esqlabsR-plot-results.Rmd
esqlabsR-plot-results.Rmd

Plotting the simulation results is an integral part of model diagnostics and quality control. esqlabsR implements an excel-based Plotting Workflow for figure creation directly from simulated scenarios. It is also possible to use esqlabsR’s functions to plot using R code.

Plotting Workflow

This process relies on filling a Plots.xlsx file in order to specify all the figures that need to be created. Then, createPlotsFromExcel() will translate the information in the excel file into plots !

First, DataCombined are defined with simulation results coming from simulated scenarios and observed data sets in the DataCombined sheet. More information in the Create a Datacombined section.

From each DataCombined, multiple plots of different types can be defined using the plotConfiguration sheet. More information in the Customizing Plots section.

Each plot to draw is defined in a plotGrid which is specified in the plotGrids sheet. A plotGrid can contain one or several plots defined in the plotConfiguration sheet and has many options to customize the layout of the plots. More information in the Drawing Plots section.

plotGrids export options like size, quality, file type etc… are available in the exportConfiguration sheet. More information in the Export Plot section.

Once all the excel sheets are setup, just call the createPlotsFromExcel() function to generate and export the plots.

Specify a DataCombined

A DataCombined is where the data to plot is stored. It can contain simulated results and/or observed data and links them together.

First, the user needs to define the name of the dataCombined. The name need to be stored in the DataCombinedName. Then, for each DataCombinedName, the user can attach simulated results (set dataType to simulated) and observed data (set dataType to observed) and give them labels. For Simulated results, the user needs to specify the name of the scenario and the output path of the simulation. For observed data, the user needs to specify the name of the dataset to use. Finally, both type of data can be linked together if they share the same value in the groups column.

Other columns available in this sheet are relative to the data transformations that can be performed by dataCombined. For more information on this, please refer to the ospsuite documentation about it here.

Customizing Plots

The default Plot.xlsx file already contains the most common columns required to customize plots. However, the user can add all the necessary columns that are needed to reach the desired result. The list of available variables can be explored in a defaultPlotConfiguration object.

Here is a sample of the available plot settings:

#>  [1] "foldLinesLegendDiagonal" "foldLinesLegend"        
#>  [3] "lloqDirection"           "displayLLOQ"            
#>  [5] "errorbarsAlpha"          "errorbarsLinetype"      
#>  [7] "errorbarsCapSize"        "errorbarsSize"          
#>  [9] "ribbonsAlpha"            "ribbonsLinetype"        
#> [11] "ribbonsSize"             "ribbonsFill"            
#> [13] "pointsAlpha"             "pointsSize"             
#> [15] "pointsShape"

You can access the full list by running DefaultPlotConfiguration$new()

For instance, if the subtitle needs to be changed, the user can add a column named subtitle and fill it with the desired value for each plot.

Leaving a cell empty will result in the default value being used.

For properties that accept a list of values (e.g. xAxisLimits), the values should be separated by a ,. If an entry itself contains a ,, enclose it between parenthesis.

Drawing Plots

plotGrids are the objects that will be used to draw the plots. They are defined in the plotGrids sheet. The user needs to define a name for each plotGrid. For single panel plot, only one plot must be listed in the plotIDs column. esqlabsR also provide a simple way to combined several plots in a multi-panel figure, in this case, the user needs to list all the plotIDs to combine in the plotIDs column, separated by a ,.

To customize the plotGrid, the user can add all the necessary columns that are defined in the PlotGridConfiguration class. Here is a sample of the available:

#>  [1] "tagMargin"                  "tagLineHeight"             
#>  [3] "tagAngle"                   "tagVerticalJustification"  
#>  [5] "tagHorizontalJustification" "tagFontFamily"             
#>  [7] "tagFontFace"                "tagSize"                   
#>  [9] "tagColor"                   "tagPosition"               
#> [11] "tagSeparator"               "tagSuffix"                 
#> [13] "tagPrefix"                  "tagLevels"                 
#> [15] "design"

Export Plots

In order to export plots to image files, the user can use the exportConfiguration sheet. The plotGrids to export must be added in the plotGridName column. Then, the output file name must be specified in the outputName column. The output format can be customize using the properties listed in the ExportConfiguration class. Here is a sample of the available:

#> [1] "heightPerRow" "dpi"          "units"        "height"       "width"

Plotting Workflow Example

Example Scenario

For the following examples, we will simulate an example scenario as described in vignette("esqlabsR") and load the corresponding observed data.

library(esqlabsR)
# Create a project configuration
projectConfiguration <- createDefaultProjectConfiguration(example_ProjectConfiguration())
# Create `ScenarioConfiguration` objects from excel files
scenarioConfigurations <- readScenarioConfigurationFromExcel(
  scenarioNames = "TestScenario",
  projectConfiguration = projectConfiguration
)
# Run scenario configuration
scenarios <- createScenarios(scenarioConfigurations = scenarioConfigurations)
simulatedScenarios <- runScenarios(
  scenarios = scenarios
)

# Load observed data
dataSheets <- "Laskin 1982.Group A"
observedData <- loadObservedData(projectConfiguration = projectConfiguration, sheets = dataSheets)

Setup the Plot.xlsx file

Then, the Plot.xlsx file is setup to define the DataCombined, customize plots, specify plotGrids and export options. The Plot.xlsx file used in this example can be found here.

use createPlotsFromExcel()

In the next step, the user calls the function createPlotsFromExcel(), passing the generated data:

plots <- createPlotsFromExcel(
  simulatedScenarios = simulatedScenarios,
  observedData = observedData,
  plotGridNames = c("Aciclovir", "Aciclovir2"),
  projectConfiguration = projectConfiguration
)

The function returns a named list of ggplot2 objects, with names being the names of the plot grids:

names(plots)
#> [1] "Aciclovir"  "Aciclovir2"
plots$Aciclovir

plots$Aciclovir2

Also, calling this function will export the plots as image files if the exportConfiguration sheet is setup correctly.

By default, the function will try to create all plots defined in the plotGrids sheet. If any of the simulation results or the observed date required by these plots cannot be found, an error is thrown. To override this behavior, e.g., to only plot the observed data without having the results simulated, change the value of the argument stopIfNotFound to FALSE. You can also specify which plots to create with the plotGridNames argument.

Plotting With Code

In some situation, the user needs to quickly draw a plot from a simulation result object using code while wanting to use the default esqlabsR theme. In this situation, using the Plots.xlsx file is not necessary. Instead, the user can directly use esqlabsR’s and ospsuite’s plotting functions to create the desired plot while preserving the esqlabsR theme.

Basics of figure creation with {ospsuite}

Simulated modeling scenarios can be passed to plotting functions from the ospsuite package to create uniformly-looking plots. To get familiar with the DataCombined class used to store matching observed and simulated data, read the Working with DataCombined class article. The article Visualizations with DataCombined covers the basics of creating supported plot types and how to customize them.

Using esqlabsR

For the following examples, we will use the same scenario as described in the Example above.

library(esqlabsR)
# Create a project configuration
projectConfiguration <- createDefaultProjectConfiguration(example_ProjectConfiguration())
# Create `ScenarioConfiguration` objects from excel files
scenarioConfigurations <- readScenarioConfigurationFromExcel(
  scenarioNames = "TestScenario",
  projectConfiguration = projectConfiguration
)
# Run scenario configuration
scenarios <- createScenarios(scenarioConfigurations = scenarioConfigurations)
simulatedScenarios <- runScenarios(
  scenarios = scenarios
)

# Load observed data
dataSheets <- "Laskin 1982.Group A"
observedData <- loadObservedData(projectConfiguration = projectConfiguration, sheets = dataSheets)

For the following steps, no Plots.xlsx file is needed. Instead, we will use the DataCombined class to store the simulated and observed data, use the ospsuite’s plotting functions to create the plots together with the esqlabsR functions to customize the plots and their export.

Create a DataCombined object

The simulation results are stored in a list returned by the runScenarios() function. Plotting and visualization are performed by storing these results, matching observed data in a DataCombined object, and passing it to plotting functions. Observed data in the form of DataSet objects are added to a DataCombined object via the addDataSets() function, and simulated results can be added by using the addSimulationResults() function. Observed and simulated data can be linked by setting the groups argument in both methods. Data sets of the same group will then be plotted together when calling plotting functions on the DataCombined object.

Let’s create a DataCombined object and populate it with data with the following code:

dataCombined <- DataCombined$new()
dataCombined$addDataSets(observedData, names = "Observed", groups = "Aciclovir")
dataCombined$addSimulationResults(simulatedScenarios$TestScenario$results,
  names = "Simulated",
  groups = "Aciclovir"
)

You can also return the DataCombined objects defined in the DataCombined sheet with the function createDataCombinedFromExcel(), read (this section)[#specify-a-datacombined] for more information.

Customize and generate plots

Customization of the generated figures - specifying title, axes ranges, axes units, the position of the legend, etc., are done through plot configurations - objects of the class DefaultPlotConfiguration. To combine multiple plots into a multi-panel figure, create a PlotGridConfiguration object, add plots to it, and plot with the plotGrid() method. Finally, to export a plot to a file (e.g., PNG or PDF), use an ExportConfiguration object.

To use configurations with a similar look and feel in the different esqLABS projects, create the configurations using the following functions:

For the list of supported properties of the PlotGirdConfiguration, refer to the reference

The next example shows how to create a multi-panel figure using the default configurations.

plotConfig <- createEsqlabsPlotConfiguration()
gridConfig <- createEsqlabsPlotGridConfiguration()

plotConfig$title <- "Time profile"
indivPlot <- plotIndividualTimeProfile(dataCombined, defaultPlotConfiguration = plotConfig)

plotConfig$title <- "Observed-vs-simulated"
obsVsSimPlot <- plotObservedVsSimulated(dataCombined, defaultPlotConfiguration = plotConfig)

plotConfig$title <- "Res-vs-time"
resVsTimePlot <- plotResidualsVsTime(dataCombined, defaultPlotConfiguration = plotConfig)

plotConfig$title <- "Res-vs-simulated"
resVsSimPlot <- plotResidualsVsSimulated(dataCombined, defaultPlotConfiguration = plotConfig)

gridConfig$addPlots(list(indivPlot, obsVsSimPlot, resVsTimePlot, resVsSimPlot))
gridConfig$title <- "All aciclovir plots"
gridPlot <- plotGrid(gridConfig)
gridPlot

Export Plots

To save the plot to a PNG file, use the ExportConfiguration. Make sure that the fileName argument ends with .png:

exportConfig <- createEsqlabsExportConfiguration(projectConfiguration)
exportConfig$savePlot(gridPlot, fileName = "All plots.png")

By default, the height of the output figure is calculated from the number of rows in the multi-panel plot and the height defined in ExportConfiguration$heightPerRow. If you want to define a fixed height with the parameter ExportConfiguration$height, set ExportConfiguration$heightPerRow = NULL.

Observed Data

Functionalities of esqlabsR require observed data to be present as DataSet objects. Please refer to the article Observed data for information on loading data from Excel or *.pkml files. esqlabsR offers a convenience function loadObservedData() that facilitates loading data in esqLABS projects. Assuming the standard project folder structure is followed, and a valid ProjectConfiguration (see Standard workflow) and Excel files with observed data are present in the projectConfiguration$dataFolder folder, the following code loads the data:

projectConfiguration <- createDefaultProjectConfiguration(example_ProjectConfiguration())

dataSheets <- "Laskin 1982.Group A"

observedData <- loadObservedData(
  projectConfiguration = projectConfiguration,
  sheets = dataSheets
)

print(names(observedData))
#> [1] "Laskin 1982.Group A_Aciclovir_1_Human_PeripheralVenousBlood_Plasma_2.5 mg/kg_iv_"

The function loads the data from the file projectConfiguration$dataFile from the folder projectConfiguration$dataFolder and returns a list of DataSetobjects. The resulting object can be used to plot results to compare simulated and observed data.

Troubleshooting

  • At any time, you can check the groups assigned to the datasets in the DataCombined object by calling the DataCombined$groupMap or by examining the output of dataCombined$toDataFrame().

More detailed information on function signatures can be found in the following: