(US & Canada) VIDEO | Analytics Mesh Lets Users Play With Data But With Governance — AtScale VP of Product
Elif Tutuk, VP of Product at AtScale, speaks with Robert Lutton, VP at Sandhill Consultants and Editorial Board Vice Chair at CDO Magazine, in a video interview about the major shifts in the data and analytics, the hub-and-spoke approach to analytics, and the various key aspects of leveraging analytics mesh.
At the onset, Tutuk introduces Atscale as a Universal Semantic Layer platform and uses a gravity analogy to explain semantic layers. She states that two major gravity shifts have taken place within the data and analytics space:
Delving further, she explains that with the data gravity shift, all the on-prem data has been moved to the cloud to be stored in cloud data warehouses. The insight gravity shift boils down to leveraging multiple AI and BI tools.
Additionally, she mentions a third gravity shift known as knowledge gravity that sits between data and insight. In this sphere, data is transformed and made business-ready, and this is where the semantic layer resides, says Tutuk.
The organizations define business logic, transformation, and metrics so that BI and AI consumption can happen from the same place, providing a single version of truth, she says.
Moving forward, Tutuk broadly discusses the distributed nature of analytics with an emphasis on agility, governance, scalability, and integration. Speaking on the governance aspect, she mentions seeing two approaches.
The first approach is to have a decentralized data and analytics team, which focuses on implementing governance measures and definitions. Tutuk states that this approach creates a bottleneck due to its inability to keep up with the business speed.
The second approach involves having a centralized analytics team, which also has its advantages and disadvantages. Tutuk believes in the hub-and-spoke approach to governance that provides agility, governance, and integration.
She maintains that the hub-and-spoke approach not only takes the best of centralized governance but also creates federated governance enabling business units to work easily with data and create analytic products.
Highlighting organizational strategies, Tutuk discusses analytics mesh strategy. She states that just like the ongoing discussion around data products and data mesh, organizations need to pivot their thinking towards analytics products.
According to Tutuk, analytics products generate insights that are consumed by business users. Therefore, she proposes a more intensive discussion on analytics products as organizational strategies.
Elif Tutuk | VP of Product at AtScale
In continuation, she says that analytics mesh is similar to the data mesh framework, but it provides composable business definitions of data. Tutuk adds that due to the business agility factor, the definitions may change, once the objects become composable, reusable, and versionable, they can be traced.
So, an analytics mesh is a framework of all the composable semantic objects that the business units can actually reuse and then put together. Tutuk states that it has been a dream to have composable semantic definitions that can be put together to build new analytics products with governance in place.
When asked about tools needed to create an analytics mesh, Tutuk uses a Lego reference. The name originated from Danish “leg godt,” which means to play well. She states that similarly, the idea behind the analytics mesh is to let the users play well with data but with governance.
Therefore, when it comes to tools, she urges organizations to think about versioning and approaching composable semantic objects as code. Tutuk notes that analytics mesh should be able to apply those object-oriented frameworks to the semantic objects.
Further, she advises organizations to apply the best practices of software development to those objects to make them versionable with CI/CD (Continuous Integration/Continuous Delivery) integration and workflow management.
In addition, she affirms that organizations must have the technology to enable data mesh. She continues that Atscale, through its universal semantic layer, provides federated governance with high-speed cloud-optimized analytics consumption.
Emphasizing challenges, Tutuk states that they can be resolved with technology. For instance, she says, users want to use their tool of choice. Therefore, the universal semantic layer should provide a seamless integration to all AI and BI tools for broadened consumption.
The next challenge is that various data personas creating the knowledge gravity, have their ways of developing analytics products, which could be code-based or a graphical user interface. Tutuk asserts that technology should welcome the challenges and provide the best user experience.
Furthermore, she shares that AtScale welcomes BI personas as it has a graphical user interface. She also mentions working on a scale markup language and announces its release which would enable all analytic personas to create analytic meshes.
In conclusion, Tutuk states that Atscale was built with the intent of seamless BI integration. Also, she confirms it is the only platform to provide analysis on top of cloud data, to cater to multidimensional analytics at scale without extracting any data from the cloud.
CDO Magazine appreciates Elif Tutuk for sharing her insights with our global community.