Data / Machine Learning

BI Leader Qlik Goes AI, Acquires AutoML Provider Big Squid

8 Oct 2021 11:46am, by and
Qlik logo

BI and data management indie powerhouse Qlik continues to build out its solution set, moving further into the realm of AI by adding new machine learning capabilities.

In fact, Qlick recently introduced another in a series of acquisitions, expanding its franchise even more widely across the analytics lifecycle. Qlik said it has acquired Salt Lake City, Utah-based Big Squid, a provider of a standalone automated machine learning (AutoML) platform specifically designed to work with analytics and business intelligence (BI) platforms. Big Squid’s platform, while initially available as a standalone service, will ultimately be integrated into Qlik Sense, to allow enterprise BI professionals — not just data science teams — to do machine learning work.

Qlik had already offered a range of AI capabilities — including advanced analytics integration (i.e. direct integration of R and Python code, as well as API-based integration of third-party AI engines like Amazon SageMaker) and Insight Advisor, an AI assistant — that can surface insights and make suggestions. But the Big Squid application will enable Qlik to add true AutoML functionality, allowing BI pros to build models against their data sets or feed data sets into the models, to predict values that can then be integrated as additional columns.

Expanding Portfolio

Qlik’s business and acquisition strategy has been driven by Qlik’s goal of taking on the full data analytics lifecycle. With technology acquired from Attunity (now Qlik Data Integration), users can bring data in using Change Data Capture (CDC), then use technology acquired from Podium Data (now Qlik Catalog) to document the lineage of that data and perform interactive analytics in Qlik Sense.

Now users will be able to layer AutoML atop all of those capabilities and create machine learning models automatically. Important functionality — including key driver analysis at detail and aggregate levels for specific action and strategic planning, respectively; predictive analytics to forecast outcomes and take data-driven actions; and what-if scenarios to understand the impact of changes — will be offered through the integration of Big Squid into the Qlik platform.

Explainable AI, Sans Code

Big Squid will be rebranded as Qlik AutoML. It simplifies the utilization of AI in business analytics by helping users to generate machine learning models without coding. The software automatically builds models by selecting an ML algorithm — and hyperparameter values for it — that produce an optimal model. The platform also explains its choices to enable iterative model refinement for optimal results.

Shapley values will be provided for model explainability, alleviating the “black box” phenomenon that ML models often fall prey to. And while other platforms have supported Shapley values-based explainability, the Qlik/Big Squid platform provides it without the need to write Python code. Instead, users simply generate their model and can review visualizations of impactful features in the model. This is especially good for analytics teams and business users.

Josh Good, Qlik’s vice president of product marketing, told The New Stack “In the Big Squid product, the Shapley values tell you the key driver and so forth. Where I find it really compelling, once you bring it into Qlik, you get interactivity in Qlik with all the other data around it.” This interactivity enables users to see and understand the logic behind model predictions, helping users’ confidence in the model and driving their conviction to take the corresponding action.

For Analytics Teams, and ‘Citizens’

Again, the main focus of the Big Squid acquisition is to enable analytics teams to utilize AutoML. To that end, Nick Magnuson, Big Squid’s CEO, said, “Since the beginning, our product was built with the purposeful intent to empower analytics teams to do machine learning on a self-serve basis as opposed to enabling data science teams. So everything, from how our workflow operates within the product, how the UI was built, the entire UX, and the different pieces of the tech stack that we integrate directly into, was made with the intent to enable data analysts, business analysts, data engineers to build their own machine learning models and then deploy them back into the stack to use.”

But Qlik believes that AI can extend beyond analytics teams and reach everyday business users — the coveted “citizen data scientists” — too. Qlik’s Good commented, “…maybe an analytics team would set up the Qlik Sense app but it’s very realistic to see someone like a deli manager looking at that saying, ‘Hey listen, when I position my deli in a different way or on this day, I sell more of Bologna and when it rains I sell more Bologna,’ and then they are seeing that key driver come through the data and change the patterns, and how they order or arrange their store. It really is something I think we can bring to analytics teams and the people they serve.”

AI and BI

While AI is often thought of as a tool for predictive analysis, it can also be used retrospectively, just as is typically the case with BI. Though AI and BI have largely evolved separately, with new integrations like that of Qlik and Big Squid, their powers can be combined. This opens a very approachable pathway for users from different backgrounds to utilize explainable AI in business analytics, then make predictions and take action based on the insights derived.

A newsletter digest of the week’s most important stories & analyses.