Welcome to the second half ofour orientation on Spotfire Administration. In this presentation, I ll give a quick overview of the products that can be used to enhance a Spotfire environment: TIBCO Metrics, Spotfire Automation Services, Spotfire Predictive Analytics, TIBCO GeoAnalytics, and Spotfire Event Analytics. To get the most outof this half of the orientation, you should make sure to watch Part 1 first. 1
TIBCO Metrics allows organizations to create Key Performance Indicators known as KPIs based on their enterprise data. KPIs can be viewed using a browser or the TIBCO Metrics app, which is available for ios, Android, and Windows devices. KPIs present enterprise data in a simple graph. Colors red, yellow, and green are used to indicate if the indicator is on target. They can be automatically updated on a configured schedule. At this time, Metrics can make connections to Microsoft Analysis Services, Microsoft Excel, Google Spreadsheet, Oracle, Spotfire data sources that have been saved in the Spotfire Library, Microsoft SQL Server, and Teradata. For current information on the supported data sources, see the Metrics Modeler User s Guide, Metrics Modeler Tasks, Metrics Modeler Data Sources. Metrics can be purchased in conjunction with Spotfire or as a stand-alone product. 2
Behind the scenes, the core of the TIBCO Metrics environment is the Metrics Services Server. The server runs under Microsoft Internet Information Services and consists of two web applications: Metrics Viewer Service and Metrics Application Service. Metrics Viewer Service is the application that serves KPIs to browser and mobile users. Metrics Modeler users connect to the Metrics Application Service to create and edit KPIs. The Metrics server uses a third-party LDAP server for authentication. It also requires a database for storing meta-data. The recommended platforms for the server are Windows Server 2008 R2 or Windows Server 2012. The supported database type for the Metrics database is Microsoft SQL Server. For complete up-to-date information on the system requirements, please see the TIBCO Metrics System requirements page. 3
Spotfire Automation Services is used to execute automated processes on the Spotfire Server. The first step in automating a process is to create a job using the Job Builder, which is installed as an extension of the Spotfire Analyst client. Jobs are made up of a series of tasks. In the this example, the job consists of three tasks: opening an analysis, which will automatically update its data, exporting one of the visualizations as an image, and sending an email with the exported image and a link to the full analysis. A number of common task types are provided out of the box. Custom tasks can also be added to on-premise Spotfire installations. Jobs are saved as an XML file. 4
The Automation Services Server is a web service run under Microsoft IIS as an application named SpotfireAutomation. Jobs are run by sending the job xml file to the web service. Any program that can call a web service can be used to send the job, paired with any program that can be configured to trigger it. The most common configuration is to use the Client Job Sender tool, provided by Spotfire, triggered by the Windows Task Scheduler. When Automation Services is called to run a job, it logs in to the Spotfire Server as a Spotfire user and executes the tasks. If any jobs are configured to send email, Automation Services will also need access to an SMTP server. 5
Spotfire Predictive Analytics consists of two technologies: TIBCO Enterprise Runtime for R, or TERR, and TIBCO Spotfire Statistics Services. TERR is an environment for running scripts written in R, a widely-used statistical programming language. While there are open-source R environments available, TERR offers better performance, memory management, and scalability. Statistics Services provides centralized, server-based access to statistical engines, with robust clustering, load balancing, and job queuing. The supported engine types are TERR, S+, open-source R, SAS, and MATLAB. Spotfire provides example functions for each of the supported languages. 6
Spotfire Analyst users have several TERR data functions available to them out of the box: linear regression and regression trees, and logistic regression and classification trees. Statistical developers can add more in any of the supported languages mentioned on the previous slide: TERR, S+, open-source R, SAS, or MATLAB. Developers can write function scripts directly in Spotfire Analyst, or they can use their preferred IDE, such as Eclipse, RStudio, or TIBCO Spotfire S+ Workbench. Wherever the function script is written, it will need to be registered through Spotfire Analyst, which will save it in the Spotfire Library. Analysts can then embed a copy of the function in their analyses. 7
When an analysis that includes data functions is loaded in Spotfire Analyst, the functions will be run in Analyst s local TERR environment if possible. If it s not possible to run the script locally, Analyst will connect with the appropriate Statistics Services cluster. Statistics Services will queue the job and run it using the next available engine from the engine pool. If you want to be able to queue jobs, Statistics Services requires access to a database. The supported database types are H2, Oracle, Microsoft SQL Server, and MySQL. For data functions to run in the browser-based clients, Web Player Server must have access to Statistics Services. Web Player Server has no local ability to run statistical scripts. 8
TIBCO GeoAnalytics is a cloud-based Software-as-a-Service offering. It provides geocoding and other map-related functionality. Geocoding, in case you are not familiar with the term, is the process of determining latitude and longitude coordinates for a location based on information such as the city or address. TIBCO GeoAnalytics can combine maps and other geographic data from a wide variety of sources with both internal data, such as store locations, and external data for example, weather feeds or demographic information. Its provides web service APIs that allow applications to integrate sophisticated and detailed geographic capabilities. 9
TIBCO GeoAnalytics can be used to enhance geographical analyses in Spotfire. For instance, Spotfire provides geocoding functionality down to the city level (or in some area, the zip or postal code), but not down to the street or address level. Spotfire users can import an Excel spreadsheet with addresses to be geocoded to TIBCO GeoAnalytics, then export the data with the exact latitude and longitude coordinates. This spreadsheet can be imported into Spotfire, which will use the coordinates to plot the locations. Taking it a step further, custom TERR functions or IronPython scripts can be written to use the GeoAnalytics web service APIs to present more complex data, such as an analysis of optimized driving routes. 10
Spotfire Event Analytics is a bundle of products that can be added to a Spotfire implementation. At the core is a limited license for the TIBCO StreamBase Complex Event Processing platform. StreamBase CEP can combine data from real-time streams with historical data. With the help of TERR, it can use this combined data to create predictive models. If the model meets the certain criteria for instance, if a threshold is forecast to be exceeded StreamBase CEP can pass data to and execute a pre-configured Spotfire Automation Services job. As discussed earlier, this job could be used to update and distribute a relevant analysis. For a look at Spotfire Event Analytics in action, see the Demo Gallery presentation Forecast Employee Expenses. For more details on the licensing of Spotfire Event Analytics, see the TIBCO Bundle Bills of Material page and scroll down to the TIBCO Spotfire Event Analytics Add-on Bundle heading. 11
This concludes the Spotfire Administration Orientation. 12