Analytic Development Languages Supported by Data Science

Development Scripting Languages and Custom Extensions

Spotfire Data Science tools offer a wide range of native features and possibilities where a user does not need coding at all. Nevertheless, in addition to no code options there is  also possibility to use other scripting languages for implementing data science computations. Spotfire Data Science tools typically use nodes/operators (basic elements from which the analytical process is built) as part of their graphical workflows where a user can incorporate code from various scripting languages. During the runtime, such node/operator executes the code according to the type of code (e.g. call another execution environment) and typically brings back the results which can be utilized in further analysis by consequent nodes/operators.

An example of such a workflow in Spotfire Statistica is shown below.

Spotfire Statistica® offers integration with these languages: R, Python, Spark, C#, C, Statistica Visual Basic

Spotfire® Data Science - Team Studio offers usage of these languages: R, Python, PySpark, Scala, SQL, Pig, Hive QL, Spark

Deployment (Code Generation Languages)

Predictive models (from Spotfire Statistica®) generated in C, C++, C#, Java, PMML, SAS, SQL Stored Procedure in C#, SQL User Defined Function in C#, Statistica Visual Basic and (from Spotfire® Data Science - Team Studio) generated in PMML, PFA, Spark.

Once the predictive modeling is done, tools can produce a code of the model which can be afterward used for further scoring  in Spotfire Statistica® platform itself or in other applications (deployment environments, real-time scoring engines, or even gateways).  

Execution Environments

Spotfire Data Science tools can use and invoke as part of the computational process following environments:

TERR, R, Python, most RDBMS, most flavors of Hadoop, Hive, Spark, Flogo

Analytic Market Places

Azure ML, AWS, Apervita, Algorithmia, H20, Microsoft CNTK, TIBCO Community Exchange

External models, methods, and know-how can be also taken from external sources like marketplaces. Again, you can use nodes in your analytical workflow to invoke and use information from an external source. Such nodes can be a model, a single method, or an entire analytical procedure. All of this is combined inside a single processing environment of Data Science (more precisely Spotfire Statistica®).

References

You can find some of the references connected with the topic below: 

• (video) Open source integration in Statistica
• (video) Model deployment in Statistica
• (video) H2O and other marketplaces connection with Statistica
• (video) Spart integration with Statistica
• (video, article) Integration with analytic marketplaces in Statistica
• (exchange) Community Exchange templates and extensions
• (documentation) Jupyter Notebooks in Team Studio
• (documentation) Custom operators in Team Studio
• (documentation) SQL Execute operator in Team Studio
• (documentation) Pig Execute operator in Team Studio
• (documentation) HQL Execute operator in Team Studio
• (documentation) Model export formats in Team Studio

Using development scripting languages directly from Spotfire®:

• (article) Spotfire Data Function library
• (article) Statistica data function
• (article) TERR data function