Alteryx Rolls Out AutoML, Cloud Analytics Innovations

Alteryx showed off its latest deep investments in self-service analytics and data science automation at its recent Virtual Global Inspire conference. IDN spoke with Alteryx vice president Jay Henderson about latest support for cloud and AutoML.

Tags: AI/ML, AutoML, Alteryx, analytics, cloud, data science,

Jay Henderson, Alteryx
Jay Henderson
vp, product management

"We're bringing the core capabilities of Alteryx to the cloud, extending access to more analysts, data scientists, and knowledge workers."

Intelligent Data
Analytics, Apps & Data for Success in the Digital Enterprise

July 22, 2021
Online Conference

Alteryx continues to build out innovations for a portfolio of solutions to help companies more easily adopt analytics, AI/ML and automation for data science across a broad range of skill levels. 


The latest offerings, which debuted at the Alteryx Virtual Global Inspire conference, aim to  deliver on the company’s commitment to provide analytics technologies to help workers upskill quickly and generate impactful outcomes from data, according to Alteryx execs.


Alteryx CEO Mark Anderson noted, "We are relentlessly focused on making analytics and data science automated and easy for everyone."


Suresh Vittal, chief product officer of Alteryx, added, "We are investing deeply in analytics and data science automation in the cloud -- starting with Designer in the Cloud and Alteryx Machine Learning and AI introduced today. We continue to be laser focused on being the best at democratizing analytics so millions of people can leverage the power of data." 


To learn more about the enhancements, IDN spoke in-depth with Jay Henderson, Alteryx’s vice president of product management.


"We're bringing the core capabilities of Alteryx to the cloud, extending access to more analysts, data scientists, and knowledge workers, making it even easier to share and collaborate across their organization,” he said.


In particular, IDN talked with Henderson about the Alteryx Designer Cloud and the company’s approach to automation and AI/ML.


Alteryx Designer Cloud   This brings Alteryx's popular on-prem Designer product to the cloud, enabling ubiquitous access. Designer Cloud looks to bring Alteryx Designer's most valuable and popular features to the cloud via easy, one-click browser access. 


This also means users will gain "drag-and-drop, no-code access to cloud platforms, applications and on-premise data," Henderson added, along with visual data prep, data blending, and analytic outcomes that can be published to diverse endpoints.   

Because Designer Cloud is browser-based, our customers can build workflows from anywhere – including Macs.


Many of our customers have dozens, hundreds, thousands and even tens of thousands of Alteryx end users. Designer Cloud will dramatically simplify the installation and upgrade process for these customers. Customers are always running the latest version without needing to rely on IT maintenance or downtime.  


Across all these thousands of Alteryx users, Designer Cloud provides personalization of the user experience alongside ease-of-use.

Also notable, Designer Cloud will also "absolutely integrate" with Snowflake, Henderson said, "Just like Designer, Designer Cloud will give users access to the cloud native processing power inherent to the Snowflake platform," he said. This option would let organizations easily create a seamless hybrid / multicloud data pipeline and other opportunities to bring together Alteryx / Snowflake capabilities.  

Alteryx Machine Learning This provides customers "fully guided automated machine learning" (AutoML) alongside what Henderson called "ready-to-use machine learning models." 


With the marketplace's concept of AutoML finding its way into many aspects of the enterprise/cloud stack, we asked Henderson to define Alteryx's vision for Automated ML and how users could benefit. He shared product details and some vision.

A prime mission for all of us at Alteryx is to make artificial intelligence & Machine Learning more accessible to business analysts and knowledge workers. Our Investments in Alteryx Machine Learning are focused on making ML approachable and work seamlessly with the familiar Designer products.


Alteryx Machine Learning (AML) comes with a set of innovative models. Once the user selects the target variable they want to predict, AML will automatically generate new features with Deep Feature Synthesis (DFS) and then automatically train and test all of the models and will automatically surface the most appropriate one for the job.


We take on as much data as a client has and we then run hundreds of different analytic approaches at the data and provide guidance to the user about which model (or aggravation of analytic approaches) will provide them with the most optimized outcome. 

One early adopter said these capabilities all add up to more innovation – and opening up AI/ML possibilities to non-data scientists. 


Nathan Patrick Taylor, chief information officer of Symphony Care Network, described the benefits he sees in a statement. "Alteryx Machine Learning. . . empowers analysts to create valuable machine learning models on their own. Our data science leads can then easily validate and promote the models." 
Driving deeper into the Alteryx approach to AutoML, it aims to deliver more highly accurate predictions thanks to a new patent-pending technology the company calls Deep Feature Synthesis. 


Henderson shared details on how DFS works.

DFS is our algorithm for extracting the right explanatory variables out of large and complex datasets. This is important because without the right variables, the models don't have the signal they need to learn.


Our automated DFS will look at all the different variables the user has input and will create new, more complex features, which in turn are used to train and create more accurate and robust models. For example, given a list of transactions, DFS may automatically calculate the sum or mean for a given customer ID. If it's a date variable, we may calculate the day the of week, is it a holiday, etc. Of course, we may also do more complex things like providing a churn rating based on variables like length of contract, type of device, and average monthly usage minutes.

Alteryx AutoML also includes Automated Insight Generation, which is designed to uncover hidden factors and relationships in complex data. It also sports an 'Education Mode' and data science best practices to make ML more accessible to everyone, Henderson added. 


In addition to the above technologies, Alteryx also debuted the following enhancements to its portfolio: 


Updates to Alteryx Intelligence, which extend AutoML to leveraging a wide range of data types, including unstructured and complex data Alteryx Intelligence now sports NLP (Natural Language Processing) and Text Mining with easy drag-and-drop automation; Topic Modeling and Sentiment Analysis and even computer vision capabilities (for image-based data and OCR for complex semi-structured data) 


Updates to Alteryx Designer, the company's popular Designer product, now adds analytic automation capabilities across the data pipeline. It now sports more than 300 drag-and-drop capabilities to automate data prep, blending, analytics and data science. It also enables faster creation of "automation apps" and "macros," which can be easily packaged as analytic workflows Alteryx Designer also adds deeper integration to Snowflake, UiPath, BluePrism, Tableau, Microsoft Azure, AWS and others 


Unified Platform API + SDK   The new unified platform and SDK enables users to extend and embed Alteryx functionality into their interfaces, dashboards, and business processes. The technologies also promote more rapid development of new data connectors to cloud-based platforms and applications.  


Updates to Alteryx Community   Enhanced user experience lets customers and partners accelerate learning analytics, data science and share business outcomes with personalized recommendations and a new Data Science Portal with learnings and best practices for AutoML, AI and data science. 


Summing up the rollouts, Alteryx CPO Vittal added, ““Our innovation, including the newly announced unified platform, is hyper-focused on delivering a simple and engaging user experience with a no-code, low-code approach to speeding business outcomes for the modern enterprise.”