New Features

STATISTICA 6 - NEW FEATURES

Summary

   A radically simplified and optimized user-interface, fully customizable, with rapid “Quick” tabs that include only the most essential options (version 5 style user interface is also still supported)

   The virtually endless possibilities for customization of every aspect of the program

   The built-in, industry standard Visual Basic language (enhanced by over 10,000 custom functions), allowing novices to easily record macros and logs, and programmers to develop elaborate applications that can access every aspect of the program (as of this writing STATISTICA is the only data analysis application featuring such a complete COM-based (like Microsoft Excel) implementation of Visual Basic)

   The complete Web-enablement of all input and output (with the optional WebSTATISTICA Server)

   Dramatic improvements in speed of practically all procedures (exceeding the record breaking performance of version 5)

   The elimination of virtually all system limitations (e.g., unlimited variables, cases, variable names, labels, text values; multiple datasets can be opened and processed simultaneously in the same or different instances of STATISTICA, and much more)

   Further enhancements of graphics with new graph types and hundreds of new features (including a system for managing graph styles, and powerful interactive 3D slicing, brushing and zooming tools)

   A new, powerful database Query facility, import/export options (and, in enterprise versions, options to process remote databases “in place” without the need to import the data to the local storage device)

   New, flexible ways of handling output with powerful workbooks and report editors of unlimited capacity, with multimedia tables, full ActiveX enablement of all STATISTICA documents, and much more

   New cutting-edge statistical procedures (e.g., for data mining, modeling) and significant enhancements to the existing procedures, including the elimination of most of the remaining (in version 5) design size limitations

   and countless other enhancements - none of which will ever slow you down, because they are conveniently accessible in optional tabs and controls to be activated only when needed...

A representative list of specific features,
and illustrations of the new functionality

Note that the following list is designed for users of the earlier versions of STATISTICA and includes references to features of the previous generation of our products.

New customers will benefit more from the Detailed Product Descriptions, which do not assume familiarity with the previous versions.

The following list was compiled to provide answers to many of the very specific questions about new features that our users have been asking - and which, therefore, pertain to very different levels of detail of the description of program features: From as specific as “Can I place more than 1 scale break on one axis?” to as general as “Do you offer any new types of user-interface shortcuts?”

For that reason the following list includes items from different “levels” of the description of the program. Also, note that some items that fit very well in two categories may be repeated in two sections.

If you are looking for some very specific features, you may use the Find option in your browser to find it in this list, or you can contact our Technical Support at your nearest StatSoft office (see info@statsoft.com to find an office near you or, in the US, call (918) 749-1119).

Table of Contents

System features
(see also the General
Summary list above)

Graphics

• New flexible system of Graphics Styles that radically simplifies graph customizations; custom styles can be built hierarchically (e.g., the style “NSF Required Title Format” may consist of the styles “NSF Required Font” and “NSF Recommended Text Size”), and applied to graph components with a single click of the mouse.

• New, faster graph redrawing that further improves on version 5's record breaking technologies; version 6 graph redrawing is based on multithreaded software architecture, and complex displays will be redrawn in the background using parallel processing, yielding unparalleled responsiveness of the program

• Further optimized and simplified user interface (e.g., including smart super-buttons) to speed up graph creation and customization

• New graph types (including a new, flexible implementation of means plots with error bars, advanced types of scatterplots where data points are represented by multivariate icon displays (e.g., pie charts, profiles, or Chernoff faces), or by custom defined graphics files (BMP, WMF, PNG, or JPG), bubble plots, and countless new graph options)

• The new architecture for graph layouts allows for more optimizations, new options, and more flexible control over graph components; for example, all fixed titles can be instantly converted into custom text and repositioned; custom text can be converted into fixed titles, labels, legends, footnotes, etc.

• The new architecture (see above) allows for more flexible management of components of complex graphs, such as matrix plots, or categorized graphs, where each subgraph can be isolated (copied as a separate object) and converted into a stand alone graph

• A variety of new graph scale options, including three new scale types (Probability, Logit, and Weibull)

• Multiple (parallel) scales (e.g., a Fahrenheit scale can be plotted parallel to a Celsius scale)

• Multiple fits and multiple (custom) function plots can be created in a single graph

• Multiple confidence bands and ellipses can be plotted in one graph for the same dataset

• New graphics display and analytic options (e.g., box plots or frequency pie charts can be plotted in matrix plots, scales are automatically coordinated in compound graphs such as scatterplots with histograms, or quality control charts)

• More interactive exploratory options; for example, histogram- or box plot-categorizations (or their parameters), categorized plot subsetting, marked subset criteria, etc., can all be modified interactively after the graph has been created (and even after it has been detached from its original dataset (e.g., when it is opened on another computer)

• Comprehensive selection of customized and pre-selected formats and styles for scale values (including the full complement of Microsoft Excel value formatting options)

• Multiple “true” scale break support allowing you to perform segmentation of regions on scales (or to collapse/extend ranges of plotted values for presentation or analytic purposes)

• Extended function plotting options in all graphs with many new functions, and including support for plotting sets of parametric curves (e.g., to plot circles, ellipses, or other complex shapes)

• User defined selections of statistics can be displayed in graphs (as custom text or placed in fixed titles), such as descriptives, correlations, multiple R’s, and they are automatically recalculated when the data change (e.g., monitor the results of brushing, see below)

• Many new brushing options including interactive and animated unique 2D and 3D slicing and interactively controlled 3D brushing

• An extensive set of new options for precise control over the graph layout, proportions of axes (e.g., graph axis proportions can be set to be precisely 1:1, or 1:1.05), and graph document management, graph placement on a page, etc.

• A new comprehensive set of options to manage titles, convert fixed titles into custom text and vice versa

• Support for multi-line fixed text objects (e.g., titles) and interactive (WYSIWYG) editing of all text objects in graphs (with no need to use any control codes)

• New, more comprehensive control over shading of surfaces with options to create stunning effects (by default, the levels of shading are now tied to tickmarks on the vertical axis); options to have STATISTICA create automatically the shading palettes by creating attractive gradient transitions between user-specified defining colors

• Significantly enhanced brushing tools with many new options, new types of brushes (e.g., interactive 3D slicers, cube brushes of interactively adjustable proportions) and new types of animations

• Significantly enhanced zooming tools with options to generate subgraphs by interactively (on-screen) selecting subsets as rectangular areas or 3D cubes or slices

• The new “Means with error bars” graph type (with custom confidence intervals, etc.) is now separated and quickly accessible as one of the shortcuts (on top of the Graphs menu)

• Numerous new customization options with further refinements of the user interface, changing defaults, defining custom styles/templates, etc.

• The resolution of the fitted functions can now be controlled (useful for plotting against logarithmic scales and other applications)

• The values of fitted functions (at a user requested resolutions) can be displayed as vectors of numbers

• The new format of the Graph Data Editor allows you to switch flexibly between the raw and calculated (derived) values

• Graphs are now ActiveX compatible objects and documents, they also serve as ActiveX containers

• Graphs can be saved in more formats (including the JPG and PNG (the new equivalent of GIF) formats for Internet applications)

• Cause-and-effect diagrams (Ishikawa diagrams) are now included among process analysis options

• Customized graphic files can be used as point-markers in all graphs

• Full programmability of all graphics via Visual Basic; like all procedures in STATISTICA, all graphs are fully programmable, and practically all aspects of the graph are accessible COM objects; thus, you can create highly customized graphs with drawn objects, embedded bitmaps or other graphs, custom text, etc. via Visual Basic either from within STATISTICA or from other applications (e.g., Microsoft Excel, Word, C++, Java, C#, VB.NET); STATISTICA provides not only a highly efficient and very powerful environment for creating graphs interactively, but also represents the most versatile and powerful graphics library available that will permit licensed users to add breakthrough graphics technology to their own (non-STATISTICA) applications

Data Management, Spreadsheets

• STATISTICA spreadsheets are now based on the StatSoft “multimedia table” technology, they are even faster than the version 5 tables, and they do not impose any size limitations

• Multimedia tables can also store pictures, sound (e.g., voice annotation of cells), music, and video; custom user interface can easily be added

• New technologies and proprietary optimizations will significantly speed up database operations and processing of extremely large data files

• All data management functions are now available at any point of the analysis (there is no separate Data Management module in STATISTICA 6)

• The identical spreadsheet functionality is now offered in all input and output (former “Scrollsheet”) tables; every output table can be instantly used for input

• Cell contents can now be of unlimited size (e.g., unlimited numbers of characters), cells can consist of multiple lines of text or numbers; spreadsheets can contain unlimited multimedia contents, custom built-in user interface, etc.

• Excel-like, comprehensive custom format specification (“parser”) facility is now offered for precise control over the display formatting of values

• The new format control allows you to assign display formats not only at the variable level (i.e., all cells in one column the same) but also at the cell or block level (like in Microsoft Excel)

• The new flexible Spreadsheets Layouts facility to manage (and quickly apply) specified styles to the Spreadsheet appearance (also applicable to special cell displays, such as selection conditions, active cell, marked cells, etc.)

• Extensive in-cell formatting is now supported (more in-cell formatting options than in Microsoft Excel), including special characters such as a b Õ f l S , sub_scripts, super^scripts, line breaks, 3D effect, etc.

• When special formatting is applied to variable names, text values, case names (now all of unlimited lengths), etc., the respective fonts, font sizes, colors, etc. will be preserved in all results, e.g., used in graphs to identify variables and/or points in scatterplots

• More interactive format/layout control tools are now supported in spreadsheets (e.g., stretching, shrinking, collapsing, and un-collapsing rows and columns, various cell/area marking options)

• The options to automatically extend lists of values by expanding blocks (e.g., Jan, Feb, Mar, ...) can be now customized and user defined lists can be added

• Microsoft Excel-style spreadsheet format painter is now available (with options to apply formats to ranges of variables, collections of cells, etc., or quickly copy formats between variables)

• Vertical orientation of text in cells, various alignment options, borders, backgrounds, a full complement of predefined and custom value formats, styles/layouts

• New data (variable) types supported in spreadsheet (including space saving integer type, extremely economical 1-byte type, and true text type of unlimited length)

• New import/output options, increased speed and (now unlimited) capacity of input operations

• New import options (e.g., a new preview window, options to ignore headers of ASCII files - e.g., drop the first two lines), and new built-in powerful STATISTICA Query application with user-friendly tools to build database queries of any complexity (see below)

• Statistics functions (which intelligently adjust for missing data) are now available as spreadsheet formula functions, e.g.:

  = mean(v1:20, v25, v30:35, time1, time3)

• New expanded syntax of spreadsheet formulas, selection conditions, categorization conditions, etc.

• All import options are available via the Open dialog (or by dragging a foreign file into STATISTICA); documents of various types can be imported (e.g., Microsoft Word documents)

• All export options are available via the Save As dialog

• The “binoculars” tool on the spreadsheet tool bar makes searching quick and easy

• Flexible case selection condition marking options are supported in spreadsheets

• A flexible set of tools for drawing random samples from large datasets is now integrated with the case selection conditions

• Spreadsheets can be password protected (with highly secure data encryption)

• Custom event control can be used for Excel-style (i.e., cell-oriented), as opposed to the usual database style (column-oriented) recalculation of specific values, write protection, custom saving, printing, editing support, etc.

• Full programmability of all spreadsheet and multimedia table operations and features; like all other procedures in STATISTICA 6, all spreadsheet options and features are fully programmable, and practically all aspects of the spreadsheet are accessible COM objects; thus, you can create Visual Basic programs (or programs written within Microsoft Excel, Word, C++, Java, C#, VB.NET, etc.) to create or delete spreadsheets, to open or save them, customize them, and perform all operations on them

Workbooks

• The workbook keeps a table of contents on the left side, organized automatically in a hierarchical manner to reflect different analyses, graphs, analyses based on different input data files, etc.; folders (representing, for example, different analyses) in the workbook can be collapsed or expanded with a single mouse-click, for example to expose all input and results documents associated with the respective analysis

• Workbooks can be saved in a single document file, and reopened to reveal all results, graphs, data files, etc. as originally placed there; for example, an entire analysis project with all input data files, graphs, results of numerous analyses, formatted written reports with embedded results graphs and tables, voice annotation, etc. can be saved to a single workbook file, and sent via email to a colleague

• Workbooks are powerful ActiveX documents containers of practically unlimited capacity; all documents can be edited and processed in-place

• The hierarchical tree structure of documents (for navigation and organization) in workbooks can easily be modified, for example, by copying branches between Workbooks

• Items can be easily dragged/dropped from specific workbooks and put on the application workspace (e.g. this means one can easily put together information from different workbooks for a desired customized effect)

• You may drag and drop, cut, delete and move noncontiguous collections of objects, branches, or sets of branches in workbooks, selectively copy, save, print

• Workbooks display features are easily customizable (proportions of panes, positions and styles of tabs, etc.)

• Objects in workbooks can be dynamically linked to other objects

• Workbooks can be easily annotated (by typing into the document, or entering notes in the comments box of the 'Document Properties' dialog)

• Full programmability of all workbook operations and features; like all procedures in STATISTICA, all workbook options and features are fully programmable, and practically all aspects of the workbook are accessible COM objects; thus, you can create Visual Basic programs (or programs written within Microsoft Excel, Word, C++, Java, C#, VB.NET, etc.) to create or delete workbooks, to open or save them, to create folders and individual workbook entries, change titles of folders, etc.

System Limits Removed

User Interface

Programming - the functionality previously offered
in STATISTICA Basic and SCL (the STATISTICA
Command Language)

• All functionality of STATISTICA 6 is accessible via Visual Basic (or any compatible environment, e.g., Microsoft Excel, Word, C++, Java).

• For example, STATISTICA breakthrough graphics technology, with practically all customization and custom graphics detailing options is fully accessible from STATISTICA Visual Basic, or can be added to other (non-STATISTICA) applications

• Over 10,000 new functions added to the standard comprehensive syntax of ms VB, thus comprising one of the largest and richest development environments available

• Takes full advantage of the object model architecture of STATISTICA

• STATISTICA automatically records Visual Basic scripts of any analysis you perform; you can record long sequences of operations and later re-run them, or edit them and build them into custom applications

• STATISTICA Visual Basic includes options for custom-designing your own user-interface dialogs; by combining the available statistical analysis techniques into custom programs, new complete applications can be developed.

• Simple keyboard macros can also be recorded to automate routine operations

• STATISTICA Visual Basic comes with a professional development environment including a real debugger (with breakpoints), editor with color coding, “type ahead” features, integrated function/object browsers, dialog painter, etc.

• STATISTICA exposes all its functionality via its comprehensive COM architecture (e.g., all STATISTICA functions can be accessed from Java, C++, Microsoft Excel, C#, VB.NET, and many other environments)

• STATISTICA Visual Basic programs can easily interface with external programs (e.g., call procedures written in other programming languages, legacy programs, integrate with existing environments)

Querying

• The new and elegant STATISTICA Query designer that allows data extraction from any relational database management system: knowledge of SQL is not required

• STATISTICA Query offers power and ease of use found only in the designated database applications (however, it is not an optional module but a component of every STATISTICA product)

• STATISTICA Query uses OLE DB technology, and is more efficient and flexible than ODBC conventions (ODBC is also supported)

• The “in-place processing” of data on remote servers option (i.e., processing data remotely without having to first import them to the local storage device), offered in all enterprise versions of STATISTICA 6, offers significant performance gains and allows you to process datasets that exceed the storage capacity of the local hardware (e.g., terabytes of data)

• Full programmability of all Query operations and features; like all procedures in STATISTICA, all Query options and features are fully programmable and accessible COM objects; thus, you can create Visual Basic programs (or programs written within Microsoft Excel, Word, C++, Java, C#, VB.NET, etc.) to create custom queries and query driven operations

Web enablement

Reporting

• A new powerful report editor of unlimited capacity (with a tree view of objects allowing you to quickly locate specific output objects and their selective management, such as printing, deleting)

• The STATISTICA reports are ActiveX containers (e.g., you can insert in them Microsoft Excel or Word files and edit them in place)

• You can drag objects into the reports and then extensively edit them and customize the layout

• A toolbar button Add to Report will instantly append the current table or graph to the report

• Reports can be saved as RTF (or STR with the tree view information) files, or saved as HTML documents and sent to a Web Server

• Reports can also be placed in workbooks, along with all other results spreadsheets and graphs

• As all other features of STATISTICA, all options for editing, formatting, placing and deleting text, graph objects, spreadsheets, etc. are fully accessible and programmable (customizable) via Visual Basic (or programs written within Microsoft Excel, Word, C++, Java, C#, VB.NET, etc.)

Statistics

A number of very significant new procedures (modules) and analysis techniques, and also graphs, have been added to STATISTICA 6. The computational speed has been dramatically improved (exceeding the record breaking performance of version 5), and the capacity of practically all commonly used procedures (where limitations existed in version 5) has been increased beyond the scope of what can possibly be encountered in actual applications. The following points briefly highlight some major improvements and additions.

• Improvements to basic descriptive statistics; extremely large numbers of variables are now supported in a single analysis (e.g., correlations), complete user-defined selection of statistics in breakdown analyses (e.g., medians, percentiles, confidence limits for the means, etc.); save stacked (multiple-group) correlation matrices; plot of means and their confidence intervals; t-tests for single samples, and much more

• Improvements to nonparametric statistics; practically unlimited number of cases in rank-based statistical procedures, Kruskal-Wallis tests can now be computed for hundreds of groups

• Distribution fitting procedures can now handle extremely large datasets

• Improvements to standard multiple regression; create instantly a stand-alone input dataset with predicted values and residuals, to be used in subsequent analyses; macros with full interactive user interface for regression analyses by groups

• New Principal Components and Classification module that allows you to map supplementary variables and cases (observations) into the factor space created from an analysis sample and variables; designated plots of factor loadings and scores, to highlight supplementary variables and cases (a very powerful tool for classification, exploration and data mining)

• Improvements and new options for nonlinear regression; designated nonlinear least squares estimation module via an optimized Levenberg-Marquardt and Gauss-Newton methods; practically unlimited sizes of data files

• Dramatically improved speed and further increased capacity for GLM (General Linear Models) analyses; instantly create stand-alone input spreadsheets with predicted and residual statistics for all cases (analysis, validation, and prediction samples), to be used in subsequent analyses

• Dramatically improved speed and further increased capacity for General Regression Models (GRM); for fast best subset and stepwise analyses.

• The ANOVA module now represents a subset of the General Linear Model

• New General Discriminant Function Analysis (GDA) module, which applies the general linear model to the prediction of categorical dependent variables, and represents a very efficient general algorithm for exploratory classification and data mining; in addition to the traditional discriminant function analysis, GDA allows you to apply best-subset selection and stepwise selection of predictors, and it will handle ANOVA-like designs with categorical predictors; best subset selection of predictors can (in addition to other criteria) be based on misclassification rates in a validation sample, so this module can be used as a very efficient classifier for data mining tasks

• New Generalized Additive Models (GAM) module; includes a complete implementation of leading-edge techniques developed and popularized by Hastie and Tibshirani (1990)

• New General Classification and Regression Trees (GTrees) module offers a complete implementation of techniques for continuous and categorical dependent variables; in addition to traditional types of analyses, GTrees allows you to specify ANOVA-like predictor designs for categorical and/or continuous variables; automatic pruning and best-tree selection; unique tree-browser to compare solutions and nodes, much more; as of this writing, to the best of our knowledge, GTrees offers by far the most powerful implementation of classification and regression tree data mining techniques available on the market

• New General CHAID (Chi-square Automatic Interaction Detection) module (GCHAID); GCHAID is a complete implementation of these techniques, and one of its unique features is that it allows you to specify and analyze ANOVA-like predictor designs; also, its results are presented in a unique interactive tree browser

• A comprehensive selection of data mining procedures (from the optional, enterprise STATISTICA Data Miner application) are now accessible directly from the Statistics menu; also STATISTICA 6 enhancements to the capacity, speed, and versatility of statistical procedures (as well as its powerful Query tools, see above) make it an even more powerful tool for data mining

• A flexible random subsampling of data facility is now available on a tab in the case selection conditions dialog

• A large number of statistical features have been added to graphical procedures, including displays of relevant statistics automatically integrated into graphs and updated when the data change (e.g., in Brushing), multiple confidence intervals and confidence ellipses, and others

• The Lowess fitting procedure is now available in all X-Y type of graphs (and as a native function of STATISTICA Visual Basic that can generate tables of predicted values)

• A new shortcut procedure to calculate and plot confidence intervals for means (a Plot of Means with Error Bars) is now accessible with a single click

• Powerful new classifiers (e.g., GDA; see above) and nonparametric (non-linear) methods (e.g., GAM; see above)

• A wider selection of statistical functions is now available in spreadsheet formulas and all those instances in the program where expression interpreters are used; in spreadsheet formulas a convenient method to quickly reference ranges of variables is supported (e.g., =mean(v1:v10, v15:v20)) and those functions automatically adjust for missing data

• Enhancements of all industrial quality control methods and techniques, and designated (customizable) toolbars for Six Sigma breakthrough quality management

• New multiple graph display and predefined reports recommended by the Six Sigma methodology (e.g., a SixGraph and other summaries for quality control)

• New cause-and-effect (Ishikawa) diagrams

• New Six Sigma calculator

• STATISTICA Neural Networks 6 is now seamlessly integrated into the system with the same look and feel and direct support for all STATISTICA 6 controls; it also offers a large variety of new, leading edge analytic options

• As all other features of STATISTICA 6, all statistical procedures, options, results, etc. are accessible programmatically (e.g., via Visual Basic, either from within STATISTICA Visual Basic, or from other applications, such as Microsoft Excel, Word, C++, Java); thus custom statistical applications can be added to STATISTICA, and the methods available in STATISTICA can augment other applications

2300 East 14th Street, Tulsa, OK 74104
Phone: (918) 749-1119; Fax: (918) 749-2217
e-mail: info@statsoft.com

©Copyright StatSoft, Inc., 1984-2008.
StatSoft, StatSoft logo, and STATISTICA, are trademarks of StatSoft, Inc.