In anticipation of the long holiday weekend for Americans and Canadians, news was fairly light in the data world this week; most announcements were around updates and enhancements to existing products.
If you would like your announcement to be included in Amalgam Insights’ weekly data and analytics roundups, please email firstname.lastname@example.org.
Product Launches and Updates
On June 29, Tableau announced the release of Tableau 2021.2, with new and enhanced augmented analytics capabilities. “Ask Data,” a capability that allows users to ask business questions of their data using natural language, and “Explain Data,” a function that provides explanations of data points, both have new interfaces that enhance users’ understanding of their data. Other new features in Tableau 2021.2 include the ability to save clean data from Tableau Prep into Google BigQuery, and to implement machine learning models from Amazon SageMaker within Tableau dashboards.
TIBCO announced the release of Spotfire 11.4 LTS on June 30. Key features in this release include the ability for nontechnical users to embed advanced analytics functions into Spotfire apps, and over a dozen new custom visualizations and apps available as “Spotfire Mods” on the TIBCO Exchange.
Anaconda announced enhancements to its ongoing partnership with Intel, including better access to libraries and packages optimized for Intel hardware to enhance the performance of machine learning models. Of note, the Intel Extension for Scikit-learn is now available in Anaconda’s package repository; Anaconda says models built using the extension run 27-36x faster than models based on the baseline Scikit-learn.
On June 29, JetBrains announced the availability of Datalore Enterprise, an on-premises collaborative version of their single-user cloud-based data science platform. Datalore Enterprise will provide JetBrains collaboration tools atop Jupyter Notebooks, along with existing features of Datalore such as PyCharm coding assistance tools.
Finally, Facebook AI announced the latest version of their Habitat platform (Habitat 2.0), a simulation platform that lets AI researchers teach machines to navigate and interact with both virtual and physical 3D environments. Improvements include ReplicaCAD, an extension of Facebook’s Replica data set, built to support movement and object manipulation as a digital twin, In collaboration with Matterport, Facebook AI also published HM3D, an open-source licensed data set consisting of over 1,000 indoor 3D scans. (This last year, prospective property buyers couldn’t go to open houses, but they could at least investigate a given property’s digital twin, and Matterport supplied a number of these virtual house tours for property listings.) Future AI-enhanced assistants and robots will need to interact with complex 3D environments; advancing “embodied” AI will be a top priority in order to build such assistants. Suggested scenarios include asking one’s AI-enhanced glasses where your housekeys were last observed, or asking a robot to check your desk for your laptop and if it’s there, to bring it to you.