Overview

Spotfire is a data mining and discovery application - "siftware" - designed to allow dynamic querying, interactive analysis, and visualization of large numbers of complex data sets in an easy-to-understand environment. It uses queries and graphical visualizations to explore trends and patterns, and filtering tools to focus on the important aspects of the analysis supporting a hypothesis. Spotfire Pro provides a means to make informed decisions by extracting valuable information from large amounts of interconnected data. It also allows users to identify easily the individual elements or target compounds that fulfill a defined set of selection criteria.

How Long Did It Take to Learn to Use It Productively?

Learning the basics of the program is very easy; one becomes accustomed to it within a few hours of clicking around. Learning to use it with full effectiveness might take three or four days.

Product Quality

Customizability

Settings are highly customizable. A large set of visualization data can be reduced to few candidate elements using the supplied filtering options, called query devices. The following query devices (filters) are included with the program:

• The range slider presets the analysis range of the data to reduce the number of data elements being analyzed.
• The item slider restricts visualization of the data to items that comply with the selection rules.
• Radio buttons and check boxes are additional data filters allowing a user, for instance, to include or exclude individual data values that support a hypothesis.

Ability to Program in Scripts, Add Extension Modules, etc.

Doesn't allow direct scripting. However, all information and settings used for one analysis can be saved or exported for use with other data sets. Moreover, several add-ons enhance the product:

• Spotfire Discovery Server (SDS) is a server and collaboration tool for depositing visualizations and annotations so that they can be shared among other researchers. It can also perform preplanned analyses based on predetermined query criteria.
• Spotfire Software Lab (SSL) allows users to customize Spotfire Pro to their particular scientific applications and research needs.
• Spotfire Structure Visualizer (SSV) links Spotfire Pro to the chemical data information management tool ISIS, allowing chemical structures and properties to be correlated to the data being analyzed.

Ability to Import and Export in Different File Formats

Can import data files in ASCII text format (delimited by tabs, commas, or semicolons) or from Microsoft Excel spreadsheets. It can access external data sources (Microsoft Access, Oracle, or SQL servers) via OLE DB and ODBC connections, and also allows users to input data from other Windows applications using the clipboard. The software can export data sets in either Excel or tab-delimited text formats. Data visualizations can be published as HTML source code or printed; the software includes a primitive HTML editor and several templates for generating Web-based final presentations. Additionally, the query settings for an analysis can be exported as an SQL database query to be applied to other data sets.

Useful or Unusual Features

(1) Data variables (legend images) of an item are represented quite uniquely by Spotfire Pro, that is, the degree of variation is reflected by legend icons that vary in size, shape, fill-in coloring, and rotation: a single cross-point on a graph can potentially convey five types of parameters. (For example, a chemical may be represented by a square whose size represents its half-life, whose color represents its in vivo activity, and whose rotation conveys its efficacy. Simply glancing at the square instantly provides a user with a wealth of data about the compound.

(2) The View Tip feature enables scientists to scan attribute pairs quickly with some degree of correlation. This feature plots information-rich scattergrams or histograms of data attributes, and a scroll-through window guides users in selecting the best visualizations.

(3) Auto Tile arranges the open windows on-screen so that all windows are visible - very helpful for interacting with several visualized outputs at once and getting an overview of a data analysis.

(4) Three-dimensional visualization is useful for complex data sets.

(5) Automated creation of dynamic query devices at the beginning of an analysis session - one for each column of the data set - simplifies the analysis process.

(6) A powerful data binning (categorizing/grouping) method allows the user to extract from preexisting data an additional set of data fitting a predefined query criterion.

(7) Can read MapInfo geographical location data files (*.mif) to associate data with geographical locations.

(8) Background images in JPEG, PNG, or BMP formats can be added to an existing visualization in order to understand data in context and to enhance final reports.

Limitations

(1) Displays are not visually pleasing. Scroll bars take up a lot of space when multiple windows (query devices) are open, even on a 17-inch monitor. Display and meaning of legends could be more user-friendly. (The scroll bars can be eliminated from view to provide a larger image area.)

(2) Toolbar shortcuts for zooming and unzooming would increase usability.

(3) It would be helpful if a legend for the color assignments were visible - perhaps integrated into the title of each window so that it wouldn't take up precious screen space.

(4) More statistical data analysis tools need to be included in the Data and Background option of the Properties pop-up window. Some descriptive statistics tools are available to be added into a HTML template to obtain these measurements, but the process is not user friendly.

(5) There should be an easier way to view and report the statistical results of an analysis, such as r and r².

(6) On the item slider, coarse and fine selection tools are hard to manipulate . (This may be an inherent disadvantage of the Windows operating system.)

(7) Allows only one analysis session to be active at a time. To open another project, the analysis in progress has to be closed.

(8) An Apple Macintosh version of the software is not available. Under the Windows OS, Spotfire Pro displays look unpolished, and the program's power seems limited by Windows' inferior graphical user interface.

Comparisons with Similar Software

Spotfire Pro version 4.0 can be used to search and analyze more data from multiple sources in a single query than the previous version, 3.3. New features include full-text searching of the data points containing a given text, three-dimensional visualization of graphical outputs, and support for Microsoft ActiveX controls. Final presentation of the results as professional-looking reports is also improved in version 4.0.

There are several products that can compete with Spotfire Pro. One of the most widely used is Microsoft Excel, which can be used to visualize data in different formats. Spotfire Pro does the job more easily, and puts the user in control of the data. Moreover, Spotfire Pro analysis is dynamic and interactive. Another strong contender is SPSS's comprehensive suite of statistical and data-mining tools, including Clementine, Clementine Solution Publisher, AnswerTree, and SPSS Diamond. Paradoxically, SPSS's breadth of possibilities may actually make users feel less in control of their data. Spotfire Pro's more limited options may actually be an advantage for those needing fast exploratory data analysis and data discovery. Since data mining is presently a burgeoning area for software developers, there are a number of other options. Extensive listings of data mining tools and how to obtain information about them are available at two Web sites, the Data Warehousing Information Center Home Page and Knowledge Discovery Nuggets.

Technical Support and Documentation

The well-written manual walks the user through the steps required to analyze a set of examples from various different areas of science. It also provides supplementary information on various data mining topics. (However, the "Installation and Setup" section is illogically placed. When I tried to install the program, the first place I instinctively looked for this was the beginning of the manual, but I had no luck and had to scroll through the contents listing to locate it. It should be after "Getting Started" rather than in an appendix at the end of the manual.)

Spotfire Inc. offers online help that is useful for beginning and intermediate-level users. For more advanced users and those requiring customization, the company offers fee-based training, maintenance, and support within North America and Europe. Additional Web-based support, though not fast, can also be obtained by emailing Spotfire support. On-site support is also available to help users quickly implement Spotfire Pro to address specific research needs.

Target Users

Spotfire Pro will be useful for scientists engaged in lead identification and in high-throughput discovery research and development, where a variety of large, complex data sets are generated and analyzed. Primary users are expected to be life science researchers (molecular biologists, chemists, quality control professionals), materials scientists, and bioinformaticians. The software can also be used by management information system (MIS) professionals.