Table of Contents

PCA Graphs

Inspect3D has a window with six displays for exploring the results of Principal Component Analysis (PCA). The six displays can be navigated between using the tabs at the bottom of the PCA Graph window. Each display can be dragged to reorder or popped-out into its own standalone window by clicking on the double window icon at the top right of the PCA Graph window. Many aspects of each display are customizable, with the user able to specify line styles, colours, and axis labels among other display features.

Note that these windows will only be populated after PCA is run on the loaded data through the PCA drop down menu. For a detailed explanation on how to run PCA and a basic explanation of the interpretation see the PCA tutorial.

Displays

The PCA Graphs window has six tabbed displays, each of which provides a unique view of your PCA results.

Variance Explained

The Variance Explained display contains a bar for each principal component (PC) that was calculated. The line along the back of the graph shows the cumulative amount of variance explained. The total variance explained by all of the calculated PCs is written in the top left corner of the graph.

Hovering over a bar causes the percentage explained by that specific PC to be displayed on top of the bar. Hovering over the points on the cumulative amount line causes the PC number, the variance explained by that PC and cumulative variance explained to appear beside the point.

Loading Vector

The Loading Vector display shows the shape of the variance explained for the selected PC at each point during the plotted cycle (also called the eigenvector). This helps the user to understand if certain PCs are associated with the variance at specific portions of the gait cycle.

The user can choose to flip the graph along the y-axis.

Workspace Scores

The Workspace Scores subwindow (formerly the Subject Scores subwindow) produces a scatter plot of subject scores for two selected PCs. The subject score for a selected PC is the distance between the subject's score for the PC falls and the mean score for that PC. This scatter plot approach can help users to visualize whether specific subjects appear to cluster together in the PC representation space.

Clicking on a subject's scatter plot point will cause the point's group and CMZ file names to appear while simultaneously highlighting the associated signal in the Queried Data window.

The user can choose to plot the subjects' normalized scores instead of their raw scores.

Group Scores

The Group Scores display shows the group scores for each PC. The group mean score for each PC is plotted as a point in that group's colour and the standard error is illustrated via error bars.

Hovering over an individual scatter point shows the underlying details for that point.

Extreme Plot

The curves in the Extreme Plot display show the shape and magnitudes of the selected PC. Within the Extreme Plot display, users can select among many different plot types.

The animation starts when the user clicks Play in the bottom right corner.

Some of the plot types allow the user to input additional specifications. In these cases, the Refresh Plot allows the user to apply their settings to the plot and have it update.

PC Reconstruction

The PC Reconstruction display shows the user what the selected groups'/subjects' underlying data would look like if it were reconstructed from the selected principal components.


Manipulating Graphs

The graphs for each of the PC displays can be customized to the user's specifications. Right-clicking on any of the graphs will bring up a context menu that allows a number of options to be set. Not all options are available for all displays.

Export Graph

This option allows the user to export the graph as a .jpg. The user controls the file that the graph is saved in and the name of the .jpg with the Export Path button. The Use Current Size check box will save the image as the same size it is in the program. If the check box is unchecked the height and the width of the image can be specified in pixels. The Cancel button will close the dialogue box without and further action. The Save button will create the .jpg file.

Colors

Background Color

This option allows the user to change the color of the background. When it is selected the Select Color dialog box, presented on the right will appear.

Axes

Depending on the graph being manipulated, there will be differences in the sub menu. The sub-menu items may be:

Selection Label This option allows the user to manipulate the x and y-axis.

Lines

Scatter Plots

Bar Charts

Error Bars

Miscellaneous