Data Modeling for the Modern World


A new world of self-service BI brings with it its own issue of data chaos. When everyone is looking at the data their own way, people find different answers to the same questions.

Data modeling can provide a valuable component of BI that has been largely neglected. In a recent DBTA webinar, Shohei Narron and Anika Kuesters Smith, data analysts at Looker, discussed the company’s approach to data modeling and how it powers a data exploration environment.

In the past, data governance has been blamed for creating a bottleneck, Narron explained. In the early 2000s, technology evolved and with more freedom, it created chaos.

“Data access was granted to a handful of people who responded to individual requests,” Narron said. “Then came the early 2000s and personal computers got powerful. Legacy tools on one hand are focused on creating standards and governance but without the flexibility of self-service access, they end up causing the data bottleneck. Visualization tools, on the other hand, are great at ad hoc self-service analytics but they cause data chaos when everyone is working on their own using different datasets and business metric calculations.”

The way to strike a balance is to put guardrails on the data, Narron said. By consolidating data in a SQL database users can load everything into one database, there are no more connectors, and users can leverage investment in data infrastructure.

Using a modern code-based modeling layer allows access to field /metric definitions, provides version control, and is consistent and reliable SQL generation regardless of tech knowledge.

Next, the BI tool needs to be put on the web, Narron said. This makes sure everyone is on the same page with the latest reports and dashboards and provides device-agnostic access.

Looker’s Look ML platform can provide all of these tools, Narron noted. The platform offers a single source of truth that defines everything the same way, is reusable, and scalable.

“A lot of what we see when we start modeling data for our customers can be individualized,” Smith said. “We have a lot of examples for users and customization.”

To view this webinar, go here.



Newsletters

Subscribe to Big Data Quarterly E-Edition