Tibco scales data analytics and visualization with an eye on AI

Enterprise analytics vendor Tibco is updating its software portfolio today with a series of releases designed to help improve analysis and data visualization with enhanced scalability, data streaming and integration of artificial intelligence (AI) capabilities.

This includes an update to Spotfire 12.2 with enhanced data visualization features as well as new data functionality for developers. Tibco acquired the original Spotfire technology in 2007 for $195 million. At the time, Spotfire was positioned as a business intelligence vendor, although the technology has steadily evolved to an AI powered data analytics and visualization platform.

The Tibco ModelOps was launched in July 2022 is now being more tightly integrated with Spotfire in a bid to enable data scientists to build and deploy analytics workflows powered by AI models. Additionally, Tibco is updating its streaming data technology with a feature the company refers to as ‘dynamic learning’ that enables data analytics operations on streaming data. The Tibco platform is also going cloud-native, with the ability for organizations to deploy software on Kubernetes infrastructure.

“We have deployments now of up to 100,000 people going into the web environment to interact with Spotfire analysis, and that can use up a lot of resources,” Michael O’ Connell, Tibco chief analytics officer told VentureBeat. “So, being able to scale that out horizontally with Kubernetes is really a big advance there for us.”

Developers and data functions

With data analytics, business analysts have commonly been a core focus for Tibco.

That said, O’Connell emphasized that Tibco is also interested in helping developers that are building out more sophisticated operations and data science initiatives. In the Spotfire 12.2 update, Tibco is bringing forward the concept of ‘mods’ as a way to bring extensibility to the Spotfire platform. With the new extensibility developers can code in JavaScript inside of Spotfire to build visual analytics operations.

Going a step further, the new data functions capabilities will allow developers to more easily take custom code and turn it into ‘point and click’ analytics operations for business analysts.

The extensibility for data analytics and data visualization can go even further than custom coding due to the Spotfire Cloud Actions capabilities. Cloud Actions provide integration points with operational systems such as databases and enterprise applications to provide a bridge between data in the Tibco platform and external sources.

Dynamic learning comes to data streaming

Streaming data is increasingly critical for business operations and data analytics.

A challenge is that streaming data often comes from an open source Apache Kafka source and must be loaded into a database before it can be used for analytics or machine learning (ML).

With the new dynamic learning feature in Tibco’s streaming data technology, O’Connell said that organizations will now be able to train ML models directly from the event stream.

“In the case of dynamic learning, we’ve developed a set of methods where you can build and train any model directly on the event stream without the intermediate step of writing the data to the database,” he said. 

One particular use case where dynamic learning is useful is in manufacturing. O’Connell said that with data coming from the manufacturing floor in real time, being able to understand and react to what’s happening as it happens is critical.

The intersection of data science and ML can be found in ModelOps

The ModelOps capabilities first released on the Tibco platform in 2022 provide organizations with model management capabilities for data science and AI workflows.

Now with Spotfire 12.2, ModelOps is being more tightly integrated, enabling what O’Connell referred to as a full data science lifecycle. Organizations can now train an ML model with the Tibco data science product that can then be cataloged and managed inside of Tibco ModelOps.

Finally, Tibco Spotfire users can find and use an ML model managed by ModelOps and use it for data visualization and analytics operations.

Looking forward, Tibco will be integrating generative AI capabilities into its platforms.

“We now are working with Microsoft and Open AI in training our own models,” said O’Connell. “We’re working with GPT 3.5. Turbo edition to create an interactive chat inside of Spotfire to make suggestions for people who are writing their own data functions. “