Recent Posts

Pages: 1 2 3 [4] 5
Tips and Tutorials / Strategies for dealing with overplotting
« Last post by Ron D. on January 01, 2015, 03:47:08 AM »
Overplotting occurs when there are so many polylines in a PC plot that all you see is a mass that cannot be distinguished as individual lines. This is quite common and quite off-putting. As Alfred Inselberg, who invented the use of parallel plots for data visualization, wrote, "Do not let the picture intimidate you!" *

The very first thing I do is pick the variable of most interest in seeing relationships, select that axis (Alt-click on the name), and do an automatic range brush of that axis with the maximum number of colors (eight) as determined in the settings. This usually immediately transforms an incomprehensible plot into something that really shows overall behavior of the data. It is an astonishing transformation, really. I repeat this for other axes if needed, and then I start color brushing smaller groups of interrelated lines. The automatic "gap brushing" option applies up to 20 different color brushes to groups of lines separated by the largest gaps on an axis, and can also be useful to sort out data that exhibits important gaps.

You can start pruning away uninteresting polylines by zooming, panning, showing values at the axes on mouse-overs, and color-brushing and hiding lines of different colors. Selected (swiped) lines can then be saved to a new data file, and when this new file is opened the ranges of the axes are automatically adjusted to the new ranges of the variables, which spreads out the lines even more.

Often the interesting aspects of PC plots are not the lines but gaps between the lines! Look for outliers and small negative spaces that can differentiate groups of polylines for selection and color brushing.

Move axes into different positions, or better yet create the PC plot matrix, to view the polyline structure between different pairs of neighboring axes. Uncorrelated variables on neighboring axes creates a cluttered, random mass of line segments between them. Correlated variables on neighboring axes produce strong direct, reciprocal or enveloped line segments. By arranging axes to show correlations you will find that clutter in the plot gives way to interesting and revealing structures.

In addition, Sliver offers two ways of applying transparency (or alpha-blending) to plots, first by invoking a separate PC plot window with adjustable transparency and linewidths, and second by exporting the plots to PDF with a transparency assigned in the settings. The plots are rendered in vector form in the PDF, so zooming while viewing the PDF is very effective in seeing the inner structure of the PC plot.

* Inselberg, Alfred. Multidimensional Detective, 1997. Download from

Tips and Tutorials / Keyboard hotkeys
« Last post by Ron D. on January 01, 2015, 03:16:07 AM »
A convenient listing of hotkey assignments for different windows is provided by the Help-->Tips menu option, also accessed by Ctrl-H. In the Grand Tour and Transparent PC Plot windows you can press H to view their hotkey assignments.

Tips and Tutorials / Nonlinear axis scales
« Last post by Ron D. on January 01, 2015, 03:09:59 AM »
As is typical with parallel coordinates, Sliver plots values linearly along each axis between its minimum maximum value. Sometimes data is better visualized when plotted non-linearly, say on a logarithmic scale. To do this, you can create new data columns that are functions of the original data columns. You might do this in Excel, or you can use the DataTools-->Add Column of Calculated Values menu option. Since PC plots don't typically have tick marks, there is no need to adjust the spacing of these along the axes.

Tips and Tutorials / UPDATED: Forcing the same range on all axes
« Last post by Ron D. on January 01, 2015, 02:58:36 AM »
UPDATED: Sliver v1.4 includes a feature to manually enter a range for one or more axes.

Sometimes data is understood better when the PC plot has consistent ranges on all axes regardless of whether a particular axes has values that extend to the ends. For example, the attached PDF export shows running speeds (inverse speeds, really, of seconds per kilometer) at each split for the top 2000 runners of the 2014 Berlin Marathon (it looks much better downloaded and opened in Acrobat than just rendered in your browser). If each axis were plotted in its own range, it would be visually difficult to tell if the runners sped up or slowed down between splits. So here each axis was forced to have the same range that spans the ranges of all the variables, except the last axis which just mapped the runner's lines to the final place overall.

To do this, you need to set the same the maximum and minimum values across all variables being plotted. One way to do this automatically is to use the DataTools-->Add Rows of Min/Max Values option. This adds a row to your data file in which all chosen variables have the same overall minimum value, and another row in which they have the same overall maximum value. For any variable you do not choose, the new rows will simply contain the minimum and maximum values of that variable, so this function will have no effect on its range.

If you prefer to set the ranges manually, or if not all variables in the data file are being plotted, you can use the DataTools-->View Column Min/Max Values option to show the maximum and minimum values for each variable. Then you can manually add a row of minimum values and a row of maximum values to the data file.

In either of the two options above, the axes of the PC plot will all have this same range when the data is plotted. There will be horizontal polylines across the tops and bottoms of the axes from the two added lines. Swipe these to select them, assign them to a color brush, and then hide that brush, leaving the original data plotted on axes with identical ranges.

An alternative available in Sliver v1.2 is to save your existing session to a file using the File-->Save Session menu option, then use a text editor to change the last n lines of the file to have the same min,max values, where n is the number of axes, and then restore the session. More information about the format of the session files can be found in Appendix C of the User's Manual. This does not require any extra lines to be swiped and hidden when the data is plotted.

Tips and Tutorials / Automatic range-brushing in reverse color order
« Last post by Ron D. on January 01, 2015, 02:37:30 AM »
You can automatically color brush an axis by selecting it (hold the Alt key while clicking on the axis name) and then pressing the "r" key once or twice or selecting the menu option Axis-->Range Brush. The number of colors (2 to 8) is set in the Options-->Plot Settings menu selection. The axis is divided up evenly into the spectrum of colors, with the blue color at the bottom and the red color at the top.

But what if you want the red end of the color spectrum at the low end of an axis rather than the high end? Sometimes low values are bad. The way to do this is to first invert the axis (again select the axis using Alt-click and then press "i"), perform the range brush operation by pressing "r", and then invert the axis back to its original orientation.

Tips and Tutorials / Setting the order in which colors are plotted
« Last post by Ron D. on January 01, 2015, 02:28:44 AM »
Sometimes you want lines or points plotted in a certain order of colors. For example, when automatically range-brushing a PC plot axis, the blue end of the spectrum is plotted first and the red end is plotted last (in front), and perhaps you would at times like the blue end of the spectrum plotted in front (I have). Or perhaps you've color brushed different sets of lines and now need to put the colors into an order that makes the lines of important colors more visible.

You can manually set the plotting order of the colors after they are color brushed by selecting the menu option Brushes-->Show/Hide/Recolor Brushes. Hide all brushes and hit Apply. Then select Show and Apply for each brush in the order in which you want them plotted, where the last color brush you show will be the topmost color.

As an additional note, when the Plots-->Display Values on Mouse-Over menu option is selected, each polyline is highlighted and brought to the front of the other lines as the mouse moves across it. This disrupts the order of plotting, but the above process can restore the plotting order by color again.

Tips and Tutorials / UPDATED: Accentuating individual lines in an exported plot
« Last post by Ron D. on January 01, 2015, 01:48:16 AM »
UPDATE: Sliver v1.4 includes a feature to thicken selected lines in the PC Plot. In a PDF export, these lines will be plotted with a specific thickness you can specify in the Options-->PostScript/PDF Settings menu option. These lines are also rendered completely opaque regardless of the alpha transparency setting, and are drawn on top of the other lines for easier viewing.

Have a look at this PDF export (it's much better to download and open it rather than open it in your browser):

It's a PC plot of the Overall Place at each split for the top 2000 finishers of the 2014 Berlin Marathon. The person I made this for (Mark) is represented by the dark blue solid polyline on top of the other semi-transparent polylines. To accentuate the visibility of his trace, after the export I directly edited the PostScript file created as an intermediate step to the PDF, and then created a new PDF from this edited PS file using Acrobat Distiller.

I've also attached this edited PostScript file, which you can download and then open in a text editor. (Yes, Sliver writes every line of this file...).

I knew Mark's final place in the race, so before I exported the PC plot I located his polyline on it by selecting the menu option Plots-->Display Values on Mouse-Over, zooming into the last scale of the PC plot, and moving the mouse until the highlighted polyline displayed his final place value on that scale. Then I swiped that line to select it and color brushed it in white.

When the setting for displaying values on mouse-over is enabled, each highlighted line is brought to the front, which can mess up the plotting order of brushes, e.g., if you range brushed an axis and you want the default blue plotted first and red last. So after zooming, selecting and coloring the desired line white it's a good idea to zoom out again and use the Brushes-->Show/Hide/Recolor Brushes menu option to first hide all the brushes and then show them again one at a time in the order you want the brushes plotted. Then you can export the plot to PDF.

After the export I opened the PS file and located the following two text  lines:

1.0000 1.0000 1.0000 setrgbcolor
1.9519 7.4915 m 4.8659 7.2346 l 7.7799 6.5696 l 10.6939 6.2900 l 13.6079 5.8743 l 16.5219 5.4965 l 19.4359 4.7483 l 22.3499 3.7734 l 25.2639 3.4787 l s

The first line is the command to set the color to white, and the next line is to draw a polyline between coordinates in the PDF file. So this is Mark's trace. I moved these two text lines to later in the file after the transparency was set back to none, and I added text lines to set the linewidth of that line wider than that which was specified for the other lines at the start of the file (derived from the export settings). I also set the color to pure blue. The final text is

[ /CA 1.0 /BM /Normal /SetTransparency pdfmark
[ /ca 1.0 /BM /Normal /SetTransparency pdfmark
% Mark's Trace.
0.03 setlinewidth
0.0000 0.0000 1.0000 setrgbcolor
% 1.0000 1.0000 1.0000 setrgbcolor
1.9519 7.4915 m 4.8659 7.2346 l 7.7799 6.5696 l 10.6939 6.2900 l 13.6079 5.8743 l 16.5219 5.4965 l 19.4359 4.7483 l 22.3499 3.7734 l 25.2639 3.4787 l s

where the % at the start of a line indicates it is commented out. The fact that this polyline is listed last also means it is plotted on top of the other polylines. Opening this PS file in Acrobat Distiller (a double-click for me) then creates the PDF with Mark's trace nicely accented relative to all the others.

Tips and Tutorials / Printing exported PDFs with transparency
« Last post by Ron D. on January 01, 2015, 01:02:21 AM »
If you export one or more plots to PDF with some transparency, you may find that printing this PDF takes a very long time or it simply hangs, as there is too much processing to be done with thousands of vector lines crossing with transparency at an enormous number of intersections. A better way is to save the PDF as a PNG file, choosing 300 or 600 dpi to get a high resolution picture, and then print the picture.

Tips and Tutorials / Getting black backgrounds in PDF exports
« Last post by Ron D. on January 01, 2015, 12:57:16 AM »
It took me quite a while to find out why my exports of plots to PDF had dark gray rather than black backgrounds like the default plot backgrounds. In the Acrobat Distiller settings, under the Color options, choose to "Leave Colors Unchanged" rather than converting them to CMYK.

Tips and Tutorials / Performance tip #1
« Last post by Ron D. on January 01, 2015, 12:53:23 AM »
If you have multiple plots up at once or an animation is running and you don't need to view the main PC plot, minimize the main PC plot window and you should see a significant performance improvement.

Pages: 1 2 3 [4] 5