View and Compare Multiple 2D Data Sets

Applies to: 2D/Studio. For XYZ data see View and Compare Multiple 3D Data Sets, which is the same procedure in tiled-pane form.

This is TableCurve Studio's visualization option for comparing many XY data sets at once. All of the selected data sets are drawn in a single graph, one above the other, so that differences between them can be seen directly rather than by switching from one node to the next. The procedure is for visualization only: no data are altered, and nothing is added to the notebook.

Opening the Comparison

· View menu > View First 25 2D Data Nodes renders the first 25 two-dimensional data nodes in the notebook.

· Right-click a data source in the Explorer and choose View and Compare N 2D Data Nodes, where N is the number of 2D data nodes that source holds. This scopes the comparison to one worksheet or file.

The right-click item appears only when the source holds more than one 2D data node: comparing a single node is just a node click. A maximum of 25 data sets are rendered.

Comparing Fits and Estimations

The same tiled comparison is available one level down, for the fitted or processed children of a data node. Right-click a data source and the menu offers a compare for each kind of child it holds, with a live count: Spline Estimations, Smoothed Splines, Local Regression, Savitzky-Golay, Automated Smoothing and SG Smoothing. Each opens one pane per child node, sharing the same Format, Normalization and Adjustments controls described below.

Format

The Format group selects how the data sets are drawn. All five formats show the same sets and the same normalization; they differ only in presentation.

· Contour is the principal visualization and the format the window opens in. Each data set occupies its own band up the Data Set axis, drawn as a colour contour with the data trace overlaid on it. Reading up the plot is reading across the series: a feature that drifts from set to set appears as a diagonal in the colour field, and a set that departs from the others is visible immediately.

· 3D Gradient renders the same stack as a three-dimensional surface coloured by a gradient. 3D Shaded renders it as a surface lit by a light source, which shows relief more strongly than colour alone.

· Single Plot abandons the stack and draws every data set in one ordinary 2D X-Y plot, overlaid on common axes. This is the format for judging amplitude and position differences directly.

· Single/Contour combines the two: the contour stack in the upper Y2 axis and the overlaid single plot in the lower Y axis.

The graph toolbar carries the same three format buttons, along with Select Color Gradient, Overlay Data, Toggle Legend, Shift Patches to Align Maxima and Hide Rendering of Patch, so the format and its display can be changed there as well.

Normalization

Data sets recorded on different scales cannot be compared as they stand -- the largest simply dominates. The Normalization group rescales every set before drawing; the original data are never modified, and the choice affects the plots only:

· None compares the data directly, with no adjustment.

· Unit Area scales each set so that its total area is 1.0. Use this when the sets represent distributions and their shapes matter more than their magnitudes.

· Unit Magnitude scales each set so that its maximum is 1.0. Use this to compare shape and position when the sets differ greatly in size.

Switching normalization redraws immediately: all three scalings are prepared when the window opens, so the choice costs nothing.

Adjustments

Force Zmin=0 holds the bottom of the colour gradient at zero rather than at the lowest value present. TableCurve's gradients are built to show a very dark colour at a true zero baseline and to change colour as soon as signal is present; with a small negative minimum, that distinction is lost across the whole plot. Check this option to restore it.

Identification

With many sets drawn together, the question is usually which one is that.

· None leaves the plot unannotated.

· Use Mouse highlights the data set nearest the pointer and identifies it by its main title.

· Use Legend highlights a set when the pointer is over its legend title.

Zooming to a Region

Begin in the Contour format and drag a box around the region of interest. Because the data sets occupy bands up the Data Set axis, the box selects both a range of x and a contiguous range of sets: zooming to sets 2 through 5 within a band of x plots only those sets, over only that region. The zoom is kept when you change format, so the 3D Gradient and 3D Shaded views show the same selection. To see every set again, zoom out to include the full Data Set axis.

Wheel and Middle Button

In the 2D formats, the mouse wheel zooms about the cursor: the point under the pointer stays fixed while the view scales around it. Hold Shift to zoom in x only, Ctrl to zoom faster. Moving the pointer off the graph restores the view as it was before the wheel was used, so exploring by wheel never loses your place. The middle button does nothing in 2D, by design -- notebook wheels fire it accidentally during scrolling.

In the 3D formats the wheel rotates the surface: wheel up and down changes the Z elevation, and a tilt wheel left and right changes the XY rotation. Hold Shift to swap which angle the wheel drives, on mice without a tilt wheel, and Ctrl for coarse 10-degree steps. A middle click restores the default viewpoint. Precision touchpads send the same messages, so two-finger gestures behave identically.

Animation

In the 3D Gradient and 3D Shaded formats the surface can be rotated continuously. Use Modify 3D Animation Settings in the graph toolbar to set the angular steps and scope, and Start/Stop Animation to run it. Click the surface to pause and click again to resume. Other options are disabled while an animation is running.

OK and Cancel

OK saves the current graph configuration -- format, normalization, adjustments and view settings -- as the default the next time this comparison is invoked. Cancel closes the window without saving; nothing from this session is remembered. Neither produces a notebook item, since the procedure is for visualization only.