What Are Composable Analytics?

Scaling AI, Featured Catie Grasso

“Businesses need more advanced and flexible analytics applications to cope with increasingly complex and unpredictable business needs. Many companies have adopted embedded analytics in order to bring analytics closer to business processes and improve decision making. Microservices or cloud-based tools add agility and enable users to receive business value more quickly. With augmented analytics becoming more prevalent within solutions, it is becoming possible to package and embed more advanced analytics capabilities hand-picked from the analytics stack within the business process,” according to Gartner.*

But how exactly is that accomplished? The answer is through composable analytics, which harness low- and no-code capabilities to go beyond embedded analytics and create consumer-focused applications from already existing analytics assets. Basically, it comes down to having a data system that contains sub-components that can be selected and assembled in a multitude of ways to satisfy specific requirements on behalf of the user. 

Not following? Here’s an example. One Dataiku customer, a multinational bank and financial services company, has developed a data marketplace that people across the organization can use when they need to get answers from other datasets. Teams can build their own projects or applications on top of that data — relevant to their specific function or line of business — and can easily share their results with stakeholders or other teams. 


Composable Analytics at Work

Analysts and citizen data scientists will likely become the main composers of analytics by reusing analytics assets to create new business value, meaning that, ultimately, the applications of the future will be assembled and composed by the people that actually use them. Here are a few examples of composable analytics at work in Dataiku:

  • Dataiku and Tableau: Dataiku’s AutoML Insights extension enables Tableau users to train ML models on the fly, then visualize and interact with key model metrics inside Tableau. The ability for the two tools to talk to each other allows analysts to combine their power within applications and easily share results via dashboards or visualizations. 

  • Dataiku applications: A big part of composable analytics is reusable assets, making it less challenging for those who want to incorporate data into their everyday work to do so. Let’s say there’s a data scientist who creates a project and realizes the workflow in the project could be useful to their analyst colleagues.

    Some analysts might just need the project as a starting point to do their own work and create a duplicate. Others want to actively participate in further development of this project and create a branch. Still others simply want to apply the existing project’s workflow to their own dataset, but don’t need to understand the details of the project. This last group can use Dataiku applications to either package a project with a GUI on top (which empowers more people within an organization to leverage AI and self-service analytics) or package part of a flow into a recipe usable in the flows of other projects. 
  • From junior to advanced analysts: Whether you are just dipping your toes in the water with data analysis or are a citizen data scientist incorporating AutoML into your work, Dataiku has something to make your job faster. If you’re doing a lot of data prep, you can use Dataiku to cleanse, combine, reshape, and enrich your data in a way that’s transparent and repeatable. If you’re more advanced, you can build models from over 10 pre-build machine learning algorithms, compare model performance, and easily share results with stakeholders.

By providing low- and no-code capabilities to our customers and, in turn, those organizations offering them to their business users, we’re working in unison to make self-service analytics cultures commonplace and drive analytics adoption at scale at companies around the world. 

*Gartner, Composable Analytics Shapes the Future of Analytics Applications, Julian Sun, 9 September 2020

You May Also Like

What Is DataOps?

Read More

What Does It Take to Democratize AI? Insights From the Field

Read More

Maintaining The Human Factor in Machine Learning

Read More

Navigating Targeted Ad Bias With Responsible AI

Read More