Statistica

Statistics Software Climate

The statistics software market is a mix of retail and freeware comprising complete packages such as Statistica, Stata, SAS, SPSS, and Minitab; complemented by programming languages, database software, spreadsheets, and numerous specialized stat packages. The complete packages offer data management and statistical tools that can meet most of the needs of most organizations.

The decade-long evolution of statistics software has involved improved user-friendliness; broader technical reach; gains in speed, automation, and data size capabilities; greater and lesser software fidelity; and enhanced multiuser environments. Meanwhile, software has remained more tool-based in organization, as opposed to problem-based, and governance has improved for safeguarding data, while relatively little for data analysis. The overall trajectory points to handling more data, greater execution speeds, more automation, rich multiuser environments, and new artificial intelligence tools.

Empowering people with push-button statistics software with no corresponding increase in governance is resulting in more statistical malfeasance. In the past, users with the training necessary to use the statistics software usually had sufficient training in statistics. Famous examples of statistical malfeasance include the Space Shuttle Challenger disaster and the election prediction of 1948, Dewey Defeats Truman.

The value proposition for statistics software is straightforward:

$$\text{ROI}=\text{Productivity/Cost,}$$

yet it can be difficult to accurately appraise. Estimating the productivity should include the synergy with other software and the user. The cost should include the shelf price, software development, maintenance, and errors.

Statistical Interface

There are multiple ways to interface with Statistica, including pulling drop-down menus, executing macros in Statistica Visual Basic (SVB) that can call another 11,000 functions in addition to those of Visual Basic, calling programming languages, scheduling automatic analysis (Monitoring and Alerting Server), accessing tools and results via the web or cloud, calling it from other software such as Spotfire and Spark, and using an [Page 1607]icon-based graphical user interface (GUI) in individual workspaces or multiuser environments (groupspaces). In addition, Statistica can deploy code, can send reports or messages, and offers a comprehensive manual. A complex analytics session can look like Figure 1, with two workspaces on the left: output and interactive decision making on the right.

The GUI enables users and groups of users to build processing flows or streams that provide a big picture and in a collaborative space. This interface permits the user to reflect in graphical pictures the conceptual flow of the steps in processing, analyzing, and presenting data. Data transformations and analyses are assigned nodes or icons in a toolbar. The nodes are pasted into a workspace and connected with arrows to represent a data flow through a sequence of processing operations. Users can create custom nodes to suit their needs, and there are nodes to call external code modules, including iron python, python 3.8.3, C#, spark scala, SVB, R, and H2O (Version 14 of Statistica). This accommodates user preferences and opens the package to existing custom code and the latest innovative ideas.

...