Alteryx 2020.3 highlighted by one-click data modeling tool

With process automation an ongoing focus for Alteryx, a new no-code data modeling capability highlights the release of Alteryx 2020.3.

Alteryx, a data management vendor founded in 1997 and based in Irvine, Calif., unveiled its latest platform update in a blog post on Sept. 1, and all of the features included in the release are now generally available to customers.

Alteryx previously offered data modeling capabilities with its Assisted Modeling Tool, but new in Alteryx 2020.3 is Automatic Mode within the tool. With a single click of a mouse, users can create a machine learning pipeline that automatically determines the best algorithms, data features and data transformations to create a data model.

By adding Automatic Mode, Alteryx is targeting users without a background in data science in addition to data experts already enabled by the Assisted Modeling Tool, according to Dave Menninger, research director of data and analytics research at Ventana Research.

"They've adopted a position that you will have data science experts and people who are dabbling in data science, and they've done a good job creating a single platform that those two audiences can share," he said. "You hear several vendors talk about the no-code approach and the code approach, and you really do need to support both."

They've adopted a position that you will have data science experts and people who are dabbling in data science, and they've done a good job creating a single platform that those two audiences can share.

He added that most serious data science is still done in the coded environment. And because those who require a no-code approach generally have other responsibilities within an organization -- they're not data scientists -- Alteryx's setup enables trained data scientists to pick up on the work done in the no-code environment and take it further.

"What Alteryx has done that is smart is create a platform where you can do both approaches in one platform, making the no-code approach part of a broader analytics platform," Menninger said. "Some of the other platforms that are offering this no-code/code approach are separate tools. Alteryx is probably one of the better ones at bringing that complete spectrum into one platform."

In addition to the Automatic Mode now available in the Assisted Modeling Tool, Alteryx 2020.3 includes:

• text mining from PDFs containing multiple languages rather than just one -- Alteryx's platform supports six languages -- in order to bring in text data from such items as invoices and surveys;
• enhancements to the Snowflake Bulk Loader so that customers can now use Snowflake's internal stage in addition to its external stage in order to reduce the complexity of data transfers and potentially costs as well;
• search tools for data catalogs all located on one page so users who need to migrate between tools can do so more easily;
• a find-and-replace tool in Alteryx Designer that originated from an Alteryx Community thread that enables users to swap out words in one step;
• a bulk editing capability in Alteryx Connect that will help customers find assets in their data catalogs and make changes;
• an extension of Alteryx's system to enable users to work in their local language that allows them to convert date and time formats in tools such as DateTime, Date Time Now and Report Header, as well as the DateTimeParse and DateTimeFormat functions in the vendor's expression editor; and
• an addition to the AMP Engine introduced in Alteryx 2020.2 that places a lightning bolt indicator next to workflows powered by the AMP Engine on Alteryx Server.

In its entirety, Alteryx 2020.3 adds to the ease of use of Alteryx's platform with the addition of more process automation.

The vendor recently rebranded its suite of tools the Alteryx APA Platform, with APA an acronym for analytic process automation. And automation is one way the vendor is making its capabilities accessible to a broad audience, while also ensuring data governance and the flexibility needed to capture and react to changes.

"It's a good position in the market," Menninger said. "We need more discipline, we need more automation. It's a large part of data prep where they have experience, and they're broadening that experience across the analytics spectrum.

"They don't have the most advanced architecture, but they are a very practical tool for this end-to-end process," he added.