I recently joined The Ravit Show, an amazing community for data science & AI professionals, to chat and demonstrate the power of data products.
For those who are just learning about data products, it’s a meaning-making approach for businesses who want to be data-driven. Data products are generated from analyzing multiple data sources to create unique valuable datasets.
Not only can data products provide insight into data access, but it can also evolve with different versions, creating new data products with existing data products.
The concept of creating data and sharing it with different domains can help the organization understand how that product is being used in their organization.
The value a data product can bring to an organization
In terms of generating value from data, there are two easily comprehensible facets of data monetization.
One is how building data products can generate revenue for the organization and secondly, to distinguish which data products should not be accessed or utilized within the organization.
Value is not just about creation, but also reducing costs by enabling users to discover data products and by ensuring business groups can share and collaborate on data products.
The Data Mesh architecture required to build a data product
Going beyond data monetization and the actual details of implementation to generate value, let’s get into what Data Mesh architecture would entail. Data Mesh encourages you to have your own infrastructure for the right use cases. The idea is to connect the infrastructure and run the analytics on top.
From a Starburst perspective, we took the capabilities of Trino to create Starburst Enterprise which can connect to data sources anywhere, regardless of location.
With Starburst, cloud-native users can quickly query data without having to build complex pipelines — a key point with Data Mesh’s third pillar: self-service data infrastructure. Part of that self-service aspect is that the data products interface is designed for data producers and consumers to query, define, and consume data products.
This is why Starburst had first mover advantage with Data Mesh. As this data management strategy became mainstream, our technology was designed and ready to assist any organization interested in implementing Data Mesh from the get-go.
A common concern from many is that it appears that you have to be ready for Data Mesh or data products. That’s not the case. However, you do need to start building the basic infrastructure, cleaning your data set, so that you build your readiness progressively.
Curious with how to build data products quickly with Starburst?
Skip to 16:10 to see me demo how easily and fast you can build data products.
Frequently asked questions when creating a Data Mesh
Other commonly asked questions include: How do I define my domains? Should I change my organization? Before I am ready for Data Mesh, should I do X, Y, Z, and then should I just move the team around?
My answer to those questions is an emphatic no! Data Mesh is supposed to help data professionals be better at becoming a data-driven organization, not create more friction. Sure, disruption is necessary, but if you disrupt the entire organization in a day, that creates tension and reduces productivity. And the bottomline is that if you are ready for digital transformation, you are already ready for Data Mesh.
