Get Ahead with Tableau as Your Enterprise Analytics Solution: Use The Semantic Layer
Getting Started: With a single Tableau desktop license, or even a trial, you can be off and realizing new insights in almost no time. With Tableau, a user can connect to almost any of their enterprise data sources, both on-premise and in the cloud. By visually examining and exploring your data, you can identify real business issues, opportunities in threats in minutes to hours. Tableau has changed the game. Before Tableau, the typical approach with classic Business Intelligence solutions, even with the best tools, would take days to weeks. Tableau broke the IT Business Intelligence paradigm by reducing the time to value from days and weeks to hours.
Scaling from Few to Many: Tableau Desktop works great for the analyst searching to discover the issues and opportunities to drive the business. Tableau server adds to the solution to deliver analytic dashboards, reports, and stories to key stakeholders across the business. Leverage Tableau Server when there are at least ten consumers or analysts. Tableau Server not only gives its users the ability to engage with the developed visualizations, but they can also create their own. The content creators to Tableau Server must take caution to publish their content properly or else chaos and confusion will spread as an infectious epidemic. Care must be taken to ensure that common business rules and metrics are accomplished centrally, and only one time. If this care is not taken, at the very least, the reports will be hard to maintain when enhancing business rules because it will need to be done many times and many places. And, at worst, the information between reports could become inconsistent and therefore create confusion.
The BI Semantic Layer: Classical BI has solved the problem of information inconsistency by leveraging a semantic layer. This is a virtual or logical view of the data that abstracts the complexities and makes it easy to navigate. Additional metrics and business rules are maintained here. Dashboards and reports pull their data from the semantic layer which has already calculated the business rules and metrics. The semantic layer is vital to the success of a BI tool with a larger user community. Tableau accomplishes the semantic layer through the publication of Tableau Data Sources.
Tableau Data Sources: Tableau’s solution for the semantic layer is accomplished with components called Tableau Data Sources. The Tableau Data Source is the document that connects to the data repository, joins the data, and maintains additional calculations. Building a Tableau Data Source is done in Tableau Desktop, the same tool used to create visualizations, dashboards, and reports. Many users of Tableau do not realize they have built the Data Source through the process of creating the visualizations in the Tableau Workbook: connecting to and defining the data is the first step. One of the most common mistakes I’ve observed with Tableau Server is that the Tableau Data Source is not published separately from the Tableau Workbook, resulting in an embedded data source and making it less shareable. This has even lead many to believe that Tableau does not have a semantic layer when indeed it does.
Best Practice: The way to avoid this common mistake and best practice is to first publish the Tableau Data Source to the Tableau Server from the Desktop tool. Then, create a new Tableau Workbook document and connect to that Tableau Server Data Source. This published Data Source becomes your semantic layer. All rules and calculations can be maintained here. Multiple Tableau documents can be created, all pointing to this Tableau Data Source, and reusing the same metrics. When the same business rule or metric is used in a report, you can be sure it will return the same result. When the rules change, you can rest knowing that it only needs to be maintained in one place and not many.