How to Buy Visualization and Data Analysis Software?

Visualization and data analysis software can cost a few hundred dollars or as much as $50,000. Typically, the more discipline-specific the functions of a visualization and data analysis tool are, the more the tool costs. Developers who cater to small niche markets need higher prices to offset the lower units sold.

Most of the leading visualization and data analysis software is available for Mac, PCs and UNIX systems. VDA tools for UNIX tend to cost more - even though they may not have any more functionality - than their PC brethren. You may be able to save thousands of dollars getting your VDA tools for Windows or NT instead of UNIX and achieve comparable levels of performance. But beware - developers are quick to point out that performance is usually better under UNIX or Windows NT. For example, the difference between the Mac and UNIX versions of VoxelView is not just speed and capacity but also key features such as handling embedded geometry and other functions that are only available for UNIX.

In some cases the Windows versions may not be able to handle the larger problems that UNIX can. Ask for references from developers and resellers and try to estimate the size of your data files before you choose a product.

The biggest challenge for VDA software buyers is making heads or tails of the jargon and terminology used by software developers: What is the difference between a VDA tool and an "visually oriented analytical graphics software package." Are pie charts Visualization? The answer is: Almost every software company that makes data analysis software claims to do data visualization, but there are enormous differences between packages.

According to John Sammis, vice president at Data Description Corp., developers of DataDesk, a leading data analysis program for the Mac, visual data analysis should enable scientists, engineers and researchers to do more than display a complex physical process such as the thrust from a jet engine color coded by temperature. "Non-physical concepts can benefit from visualization techniques, too. What should a graph of say, the response of a patient to a drug treatment look like? VDA software must provide tools that give users insight into these sorts of problems."

Before an organization buys VDA software, the evaluators should call the developers or their representatives, describe the application and ask for references. Each package has features that serve its core application areas best: It's advantageous to get the best package for the job. For example the best of today's statistical analysis packages may have VDA features specific to the types of analyses statisticians are likely to perform - but these are different than, say, an electronics engineer or biometrician would require.

Most statistical analysis packages, including (but not all) Statgraphics, STATlab, Statistica/W, SYSTAT, StatView, to name some of the more popular scientific and technical statistics packages, include several graphs for Visual Data Analysis. The real-time 3D rotation of scatterplots are popular for viewing data from an unlimited number of perspectives. Scatterplot matrices, a multi-dimensional graph popular among statisticians (See the article in the statistics section), enable users to view multi-dimensional data as a cubic-array of scatterplots. Other statistical plots, like Centroids and Chernoff faces (a graph that displays information as human faces with different features), can also reveal additional insight into data.

Visualization can include much more:

• 3D rendering - can convert a wire frame schematic into a solid;
• Finite Element Analysis - a numerical method for approximating data;
• Multiple light sources - gives a perspective view from any position with true-to-life shadows;
• Animation - turn a series of graphs into a real time movie;
• Pan and zoom - change the viewpoint and navigate through a graph or image;
• Isosurfaces and slice planes - Cut through a graph or image along any coordinates and view or select the data that can be found there;
• Stream ribbons - indicate the flow or force around an object;
• Vector fields - indicate flow or force at points on a surface or grid in a graph;
• Particle advection, noise reduction, thresholding, opacity, and seed detection - image analysis techniques to change intensities of image parameters to reveal new perspectives, views or underlying objects;
• Area measurement - tools for measuring the size of parts of a graph or image;
• Point probes - data analysis tools that show the values of data that correspond to points on a graph or in an image;
• Mirroring - showing the enantiomeric or mirror-image of a graph or image;
• Annotation - adding notes to a point, or part of a graph or image.

Likewise, data analysis packages can also include:

• Mathematics, and statistics functions - for manipulating and transforming data.
• Data acquisition tools - to directly input data from experiments and sensors;
• OLE and DDE Links - to facilitate data management;
• Data Dictionaries - to manage data better;
• SQL support - to enable you to query SQL databases.

© 1996 Scitech International, Inc.  All rights reserved

Back to articles menu

Go to next article