TIBCO Spotfire® 10.4 contains improvements in data access, visual analytics, data wrangling and developer capabilities. The signature feature in Spotfire® 10.4 is native support for Google BigQuery. With the new Google BigQuery connector you unlock interactive visual data discovery on huge amounts of data using live querying and an extensive range of analytical functions. The connector uses Google's APIs, so you don't even need to install a driver, and you sign in with your Google account in your web browser. Spotfire® 10.4 also introduces native support for ComputeDB, the TIBCO in-memory analytics database. Based on Apache Spark and Apache Geode, it delivers high throughput, low latency, and high concurrency for unified analytic workloads.
As a TIBCO Cloud? Spotfire user you can now refresh data directly from your TIBCO® Data Virtualization instances in the web client. In fact, thanks to new mechanisms in the data loading framework, data can now be refreshed directly from the Data menu of the Consumer and Business Author web clients. If a data refresh fails, the previous data is kept for the failed source.
New time saving features are also available. Keyboard shortcuts and search can now be used to quickly add visualizations, and Open from the File menu is back.
Administrators and Authors will appreciate that the user action log now collects visual analytics actions and Developers that the C# API allows IronPython scripts and C# extensions to access user information.
Note that Spotfire 10.4 is a mainstream version. Fixes to critical issues discovered after the release will only be made to the most current version and to long-term supported versions. For more information on the difference between mainstream versions and long-term supported versions, see the documentation.
Back to the What's new in Spotfire landing page.
Data Access
Native Google BigQuery access
The new, native, self-service connector for Google BigQuery doesn't require a driver and enables you to push interactive queries on to the largest amounts of data. Google Analytics users will also be able to push interactive queries on page tracking event data.
The Google BigQuery connector is available in the Connect to list, next to Google Analytics.
The Google BigQuery connector has most of the available Spotfire connector features. For example:
- Custom queries
- Prompting
- Relations
- Primary key columns
- In-database push down queries
- In-memory extracts
- On-demand data loading
- Automation Services support
- Scheduled updates support
You can also wrangle data using calculated columns.
In this example, we have kept only numerical values in the sample_duration column using a regular expression and then changed the data type to Integer using cast.
It's then easy to visualize the average sample duration per state.
Native TIBCO ComputeDB access
The TIBCO ComputeDB in-memory optimized analytics database, based on Apache Spark and Apache Geode (the open-source version of GemFire), delivers high throughput, low latency, and high concurrency for unified analytic workloads. You can combine interactive and streaming analytics with artificial intelligence in a single, easy-to-manage distributed cluster.
With Spotfire, you now have self-service access to TIBCO ComputeDB using the native TIBCO ComputeDB connector. Business users can easily connect and push interactive queries directly into TIBCO ComputeDB.
TIBCO® Data Virtualization support in TIBCO Cloud Spotfire web clients
In web and mobile clients running TIBCO Cloud? Spotfire, you can now load and refresh data from TIBCO Data Virtualization.
With an Internet-facing TIBCO Data Virtualization instance, this means that you can now run both Spotfire clients, the library and TIBCO Data Virtualization on cloud.
Author the connection using TIBCO Cloud Analyst on your Windows client.
Consume and refresh data in TIBCO Cloud Spotfire web clients.
TIBCO Drivers 1.8
The Spotfire Server is shipped with ODBC drivers for Apache Spark SQL, Cassandra and MongoDB. These have been updated to the latest versions.
Data Wrangling
Option not to store on-demand data between sessions
A new option for on-demand data tables makes it possible to select whether data loaded on-demand should be stored in the analysis automatically. Previously, on-demand data was always stored but with a new preference it's now possible to switch this functionality off, globally. This is useful, for example, when you want to save storage space in the library by reducing the size of your on-demand analysis files. It's also useful when many people share the same analysis file and have access to personal slices of on-demand data. Each user will then start with empty on-demand data tables until their personal data has been loaded from the data source.
Reload data from web client toolbar
It's now possible to reload data directly from the toolbar in the Spotfire web clients. Depending on how data is loaded and whether you are in consumer or authoring mode, the refresh options differ.
When you are in Business Author mode you can refresh all linked and all stored data sources via the menu options Data > Reload Linked Data and Data > Reload All Data.
When you are in Consumer mode you can refresh linked data sources via the menu option Data > Reload Linked Data for analysis files that have not been loaded by Scheduled Updates.
This difference and limitation protects your company from having potentially 10.000's of consumers refreshing large data sets that have been stored in the analysis either to separate the analytic workload from a production database workload, or have been configured to be loaded on a (nightly) schedule for the same purpose.
This feature is also available via Spotfire APIs.
In this Business Author example we have linked data but are also able to refresh all data, including stored and embedded data.
The data loading options, defined when creating or editing the analysis, control which data sources should be possible to refresh.
Data is kept if reload fails
Spotfire has three data loading options, Always new data, New data when possible and Stored data. When working with an opened analysis, if you choose to reload linked data and the reload fails, Spotfire will keep already loaded linked data. This can happen, for example, if the data source is missing. For data tables based on multiple data sources the result could then be partially reloaded data tables.
It's also possible to skip the reload of an embedded data table that could not be reloaded and continue with the remaining data tables.
Visual Analytics
User action logging for visual analytics
The Spotfire Action Log now logs actions based on visual analytics. For example, actions like when users create visualizations, change columns or aggregations, create/delete/rename pages, filter data, export data, etc, are now logged. This enables companies to better understand their users and their visual analytics behavior. The screenshot below illustrates an analysis where one can see at what time of the day that users used filters, what kind of filters they used and the details about which users that did the filtering.
Shortcuts for visualization types
It is now possible to quickly create new visualizations using the CTRL + a number keyboard combination. CTRL+1 creates a table, CTRL+2 a cross table and then it continues in the same order as in the visualization flyout. After CTRL+9, it starts over with CTRL+SHIFT+1.
You can hover with the mouse pointer over a visualization icon in the flyout to learn about which keyboard combination that visualization type has.
Search to add a specific visualization type
It is now possible to add a visualization by searching for a visualization type such as "map", "bar chart", etc, using the Find tool.
Open from the File menu
There is now an Open choice in the File menu, which opens the content browser flyout and lets the user browse for a new analysis, data source or data file.
Developer
C# API to access user information
This addition to the C# API allows IronPython scripts or C# extensions to access user name, domain, group/role membership and other relevant information about the current user.
Recommended Comments
There are no comments to display.