Aabel 2.0 - Scientific Plotting and Data Analysis
Scientists of all sorts require advanced graphing/plotting to explore and present data. Not everything can be jammed into a program like Excel or standard spreadsheet or business graphing programs. Our Windows-using colleagues often use programs like Origin or SigmaPlot. On the Mac, plotting programs include Kaliedagraph, Igor Pro, DeltaGraph, pro Fit, ChartSmith, and Aabel.
Aabel is a powerful program along the lines of Origin and SigmaPlot, offering plotting, statistical analysis and curve fitting. To quote the documentation, it's designed for "statistical analysis with a focus on visualization, dynamic data exploration, powerful data management, and interactive scientific graphing."
After several years of development since version 1.5.x, Gigawiz has released a major update of their program Aabel. Updates include a universal binary for Intel Macs, full Unicode support, a new statistics analyzer mode, AppleScript support, new plot types, as well as more minor improvements such as import of data from dBase, Matlab, and Splus. For a full list, check out Gigawiz's What's New documentation.
There's also now a Quick Start guide, which offers quick tutorials on running the new statistics analyzer, creating a standard X/Y scatter plot, curve fitting, and customizing charts.
This review cannot possibly cover every feature in a package like Aabel. Instead, it is offered to cover some major features, some drawbacks, and to describe Aabel's unique approach to scientific data visualization and exploration.
Aabel offers many unique statistical and data exploration features. One key of the program is the "pipeline" approach to data. (This includes plotting, statistical analysis, and data filtering.) To create any sort of plot, choose "New Visualization & Statistics Pipeline". If more than one worksheet is open, you can pick one or more worksheets to include in plotting. The new plotting window offers a blank page, with a gallery of button menus for graphs along the top. Picking a particular graph type will also bring up a palette for choosing variables and some plotting options.
Once the graph appears, this approach makes it extremely easy to explore multivariate data by flipping through different variable combinations in the side palette -- the plot will update automatically.
Similarly, for contour, 3D or histogram plots, the palette offers the ability to dynamically change the number of bins. Added in this version is the ability to manually set the size range of the bins instead, although this is still lacking for contour plot levels.
Interactivity with the pipeline model also allows clicking, brushing, selecting in plot to result in selections in the worksheet itself. This also allows customizing the point markers used for these particular points via the various controls to the left of the row. These alter the visibility (for filtering), color, style of marker, and labeling options. The intent appears to allow easy identification of outlier points and general data exploration, and this works well.
Aabel offers a wide range of scientific and technical plot types, from standard bar, pie, and scatter plots, to statisical box and wisker, histogram, contour (from XYZ data or a matrix representation), to a variety of geographic and mapping plots. This version helpfully offers the button menus along the toolbar of a visualization pipeline, but also a gallery button which categorizes plots and offers textual descriptions of each type. Unfortunately, many of the descriptions point the user directly back to the PDF user's guide. Some form of hyperlink or integration between the two sources would be helpful here.
Many graphing/plotting programs also bring up a new plot in an individual window. Since Aabel uses a page-layout approach, it is easy to add multiple plots to the same page window. On the other hand, the default plot sizes may look slightly small on-screen. If you wish to enlarge a plot, you should choose the graphics select (pointer) tool, and stretch the frame as you might in a drawing program like Canvas, Illustrator, PowerPoint, or Keynote.
Of course Aabel also offers some drawing and annotation tools, including text boxes, lines, rectangles, circles, and the like. Strangely, in order to create an arrow, one must draw a line to the point, select the line with the graphical selection tool, and then set the arrow options in the toolbar.
Overall, Aabel offers a huge variety of plot customization options and very professional quality scientific graphics. Graphics exports are to PDF format, allowing smooth resizing in many Mac programs like Keynote or Pages. Unfortunately, copying or exporting as a PDF graphic makes it impossible currently to paste into Microsoft Word or PowerPoint directly. Support for exporting as PICT, PNG, TIFF, or other graphics format would be helpful in sending the graphics to colleagues or using Aabel with Microsoft products.
Not to be left out in any scientific plotting program are features for error bars, curve fitting, and statistical analysis
Once a chart has been plotted via a visualization pipeline, curve fitting can be performed. Aabel gives the option for a normal regression, or a customized regression for a partiuclar variable pair (e.g., for a custom user-defined function). The former allows linear, polynomial, and other typical regression techniques, while the latter offers a libary of various functional forms and the ability to write out a custom functional fit. Aabel also offers the option to include confidence bars for linear regressions. (This would be a welcome feature in many other programs to give error bars on the regression itself.)
Additionally, Aabel offers a full range of statistical analysis from a visualization window by the "Statistical Analyzer" mode. This includes regression techniques included through curve fitting, albeit with the option of creating a new worksheet with predicted regression values and residual errors. Also included are a wide range of standard statisical methods and you will be guided through the analysis in the analyzer window, with the option of displaying results in the window or output into the plot window itself.
Unfortunately, Aabel often doesn't present things in a completely user-friendly manner. This means that there's definitely a learning curve when initially using the program.
For example, unlike most spreadsheet applications, setting a formula for a column of data requires you to not only enter the formula in a pop-up dialog, but click the "Calculate" button. The data is not updated until you open the dialog and re-click "calculate." Unlike the dynamic pipeline structure in the rest of the program, calculated data in a worksheet will not be updated dynamically.
As another example, to set the numerical format for a column of data, there is no menu command to do this. Instead, you should click the unlabeled rectangular button above a column. (A tooltip does pop up after about 4-5 seconds explaining that this is the "variable properties" button.) Here you will see some useful bits, including a histogram of the variable and various statistical properties. In addition you can set the numeric format (e.g., scientific), the number of decimal places, etc.
Changing the properties of a graph can also be frustrating in comparison to many other Mac plotting packages. Aabel offers a variety of selection tools. Some select data points (e.g., for filtering or labeling). The regular arrow tool is a graphical selection tool used for selecting and moving parts of a graph around on the page. To change the axis options, you must use the rectangle or lasso data selection tool, then mouse to an axis edge and double-click or bring up the contextual menu. Once there, version 2.0 brings up an improved dialog with buttons/tabs dividing the formatting options into separate pieces. (Previous versions showcased all axis formatting options in one huge, complicated window, so this is a major improvement, even if it's still a non-traditional approach to a tabbed dialog.) Also, version 2.0 allows formatting the axis via a menu option, which is a welcome addition.
Another annoyance comes when trying to copy graphics from the program into another program. Currently, this copies everything visible in the visualization pipeline window. Clicking on one or more objects and copying yields the same result.
My point is not that Aabel is necessarily hard to use, just that you may find yourself needing to search through the PDF manual to find out what you need to do. If the developers made some common functionality available through multiple paths (e.g., menu items, contextual menu, etc.), this would probably help. As described above, the chart axis formatting menu option is a welcome example of this. I also wonder if the data selection and graphical selection tools could be combined in some fashion.
Aabel offers a huge variety of scientific plotting, data exploration, and statistical analysis tools. Its interface requires some learning curve. The developers would do well to provide multiple ways to access key dialogs (e.g., graph and numeric formatting settings) and spend some time getting feedback from new users to make the program more intuitive. The degree of interactivity gained via Aabel's pipeline model makes analyzing correlations in multivariate data a breeze. On the whole, it offers excellent graphic quality for plots, and a range of indespensible statistical and data exploration tools for scientists.
Aabel 2.0 sells for $445 per license, $345 for Educational users, with discounts for larger quantities. Upgrades from previous versions are $195 per license. A free 30-day demo is available with no limitations.