Lavastorm Simplifies Predictive Analytics
Lavastorm says new analytics engine capabilities enable business analysts with limited knowledge of data science to deliver predictive insights.
Many business intelligence experts, including those who made recent business intelligence and analytics predictions for 2015, predict that data analysis will continue to become more democratic this year, with business users gaining the ability to perform their own "what-if" analyses with no assistance from IT. For that to happen, however, vendors must further simplify the application layer.
Lavastorm Analytics, a Boston-based provider of data analytics software, has simplified its Lavastorm Analytics Engine, which already employed a visual, data flow-driven approach to data analysis, by adding drag-and-drop capabilities that aim to help business analysts master predictive analytics.
Offering business analysts these kinds of capabilities can help companies cope with the shortage of data scientists, said Lavastorm's CEO Drew Rockwell.
"Rather than rely on specialists – who may be limited in availability – to look at the big picture, diagnose the situation and deploy analytical processes, we enable 'ordinary' users to do all that on their own," Rockwell wrote in an email. "We provide a platform that shows them the big picture, we provide them with the knowledge necessary to assess the situation, and we equip them with streamlined tools that require minimal configuration to deploy analytics. For instance, rather than writing scripts in proprietary languages in statistical programs, Lavastorm users can drag-and-drop pre-packaged functions, specify a few data fields, and run the process. The analytical results can then be passed on for further processing, analysis, or visualization.
That does not mean the product cannot be used for more advanced analyses, he noted. "By enhancing ease of use, we have not sacrificed functionality. Advanced users can still exercise greater control and manually write scripts to implement complex processes."
Among the platform's new predictive analytics features:
- Linear regression, so analysts can calculate a line of best fit to estimate the values of a variable of interest
- Logistic regression, so analysts can calculate probabilities of binary outcomes
- K-means clustering, which enables analysts to form a user-specified number of clusters out of data sets based on user-defined criteria
- Hierarchical clustering, which facilitates forming a user-specified number of clusters out of data sets by using an iterative process of cluster merging
- Decision tree, which helps analysts predict outcomes by identifying patterns from an existing data set
Last month Lavastorm announced that its platform's extended predictive analytics functionality was powered by the Tibco Enterprise Runtime for R (TERR) engine. The two companies have been long-term partners, providing Lavastorm Analytics integrations with Tibco Spotfire data discovery and Tibco Jaspersoft business intelligence reporting.
According to Rockwell, Lavastorm will continue to add analytical components to the product and will provide ongoing webinars and other educational content to facilitate the adoption of the enhanced functionality. The educational component is key, Rockwell said.
"Our aim is to equip Lavastorm’s users with enough knowledge to deliver functional analysis in as short a time span as possible. We expose users to a wide range of analytic methods, to help them understand the methods, and to show them how those methods can be translated into action," he wrote in an email. "We plan to deliver the educational content through several channels, including compact webinars that will explain the foundations of particular analytic methods and demonstrate their application using real-world scenarios, as well as more comprehensive in-person training modules that include formal instruction, instructor engagement and hands-on practice with Lavastorm’s tools."
True understanding of the data, as well as ease-of-use, are "fairly significant hurdles" to getting business analysts to perform predictive analytics, Rockwell said, noting that Lavastorm hopes to meet the challenge with "a uniquely integrative approach" that bridges the gaps that traditionally exist between different functional areas of a business.
"Where data acquisition, business interpolation and data analytics have typically operated in silos and been implemented using separate tools, Lavastorm allows these disparate functions to be easily and transparently linked together," he wrote. "This allows 'ordinary' users to see how everything comes together and, equipped with the appropriate set of knowledge, to do more than they could before."