Francesco Puppini is a consultant in business intelligence and data warehousing. With more than 20 years of experience in this field (with companies such as Deutsche Bank, Renault, and Barclays), he is currently an R&D engineer with Astrato, a new cloud-based data analytics platform. He is also an adjunct Professor at Bologna Business School.
He believes that BI should become as easy as e-commerce. In e-commerce, whatever product is available in the shop, can be simply added to the cart. Likewise, a system of business intelligence should offer one single point of access containing all the data elements that are in scope. Whatever data element is available in the system, can be simply added to the query.
His work on the Unified Star Schema has significantly contributed to the field of data warehousing and analytics.
Unified Star Schema – A New Method of Data Modeling
In 2020, Puppini co-authored a book titled „The Unified Star Schema: An Agile and Resilient Approach to Data Warehouse and Analytics Design“ with Bill Inmon, the father of data warehousing. In this book, they introduce a new method of data warehouse design called the Unified Star Schema (USS).
The USS addresses the challenge of self-service analytics for business users. Self-service analytics exists often only on paper, and if it exists, it does not give a lot of freedom. You can study only one business process (such as sales, finance, HR…), but what about a business user wanting to analyze all those business processes at the same time, within one interface?
“I wanted to widen the scope of what business users can analyze. With the USS, they can freely follow every aspect of the company, without worrying about incorrect numbers,” explains Francesco.
And why only now? Nobody had the audacity to change the very principle of merging the data together – using not only the joins but also the union, allowing for better flexibility.
The USS is a simplified and more agile approach to data modeling than traditional star and snowflake schemas. It is based on the following principles:
Unlike the traditional Dimensional Modeling of Kimball, with the USS approach, there is no longer a need to classify the tables in Fact Tables (Facts) and Dimension Tables (Dimensions). Although the terms are still used, with this new approach the following situations are allowed:
These situations are strictly forbidden by Kimball, however, they do happen in real life.
When using this approach, companies no longer need to adapt their data model to the constraints imposed by the Kimball methodology. This saves a lot of time and effort because data is accepted as it is, natively. Some transformations are still needed, as it is expected to be, but the workload is heavily reduced.
Francesco Puppini will be a guest at Revolt BI’s Datapunkers meeting, taking place on November 13th at 6:00 PM at the Kino Přítomnost in Prague. At this event titled “Fathers of modern data warehouse,” he will appear alongside Bill Inmon, the co-author of the Unified Star Schema and the father of data warehousing.
You can purchase tickets here.