Statistica

Statistica is far less well-known than its major competitors. This is for several reasons but most notably because of a) an historic approach to marketing was by word of mouth and b) the fact that the company has especially focused on process monitoring and manufacturing. The company is the leading supplier of enterprise-wide quality control and improvement software systems. However, while the product certainly has particular strengths in these areas it is equally at home in all types of analytic applications and across all industry sectors, with the former ranging from fraud detection to predictive maintenance and customer analytics.

An interesting consequence of the Dell acquisition was that the company migrated 300 SAS models (Dell was previously an SAS user) to the Statistica platform. The company is currently considering how to best market this migration capability.

While the company has its own direct salesforce it also makes extensive use of partners, especially in EMEA.

Statistica claims over a million users world-wide and these are spread across many industries. Customers include 3M, Pepsi, Schlumberger, Lenovo, Alstom, Borden and many others. The product is also widely used within the research community: especially within academic environments and in healthcare. Around 30% of Statistica deployments are in research institutes of one type or another, 60% in commercial organisations and 10% in government agencies.

Statistica provides an MPP (massively parallel processing) environment for building and deploying analytic and statistical models. Licenses are available for single and multi-user systems as well as enterprise licenses. Editions are available in a variety of languages including those using both Cyrillic and Chinese characters. The software runs on Windows and the standard language (R and Python are also supported) is Statistica Visual Basic (SVB). This has more than 14,000 functions. SVB also provides access to Visual Basic for Applications and the .NET framework. SVB can be recorded within Statistica and saved as a macro so that it can be re-run later. Statistica objects can be called directly from many other applications or COM-compliant programming languages (such as C#, C++, Java and VB.NET).

The product includes self-service data preparation, a wide range of functions for developing analytic models, model management (including the ability to manage third party models), natural language processing (for text mining in various languages), in-database analytics, real-time scoring, a built-in rules engine, support for languages such as R and Python as well as PMML (predictive modelling mark-up language), three visualisation engines and a large number of connectors (both to a variety of data sources and to other tools).

In the latest release (13.1) there are new network analytic capabilities that use "graphical association maps" based on an embedded graph database to automatically build models. Users can also visually explore complex inter-relationships. This will be useful for fraud analytics, influencer analytics and so forth. Also in this release, the product's in-database analytics processing has been extended from Microsoft SQL Server to include Apache Hive (on Spark), MySQL, Oracle, and Teradata.

Finally, Statistica has a close relationship with Dell Boomi (which will continue after the acquisition of Dell Software is complete). When used in conjunction you can deploy "analytic" atoms, where atoms are (Boomi) run-time engines that execute integration processes, but in this case execute analytic processes. What this means is that, if you want to, you can have a consistent analytic environment across an entire Internet of Things implementation with Statistica running at both the edge, in gateways and in the centre.

The company offers a range of on-site and remote implementation, data scientist consulting, configuration and end-user training services. There are also various on-line services provided, including video tutorials, technical documentation and a user forum.