Are Your Tableau Data Sources Ready for the New Year?
Posted On
Each year around this time, my teammates and I meet with education customers to discuss their data strategies for the upcoming year. There are some common themes that we typically cover, so for those of you in education who are thinking about how to prepare for the incoming data requests of a new school year, this is for you!
Understand your Data Usage
Tableau Server is a powerful part of the Tableau platform that encourages anyone, no matter their analytical aptitude, to dive into data they know and explore what that data is telling them. Any Server Admin or Project Owner trying to get the most return on their Tableau deployment should be regularly analyzing the content living in their Tableau Server and answering a few questions:
- How many data sources do we have?
- How many workbooks are tied to those data sources?
- How well can our users answer questions with our data?
- How much time is spent internally looking for the right data source or workbook?
Understand your Data Sources
Tableau Server has an amazing component, which ships free with the product, called Data Server. This small and often overlooked feature is the first step for customers to identify and communicate why data exists on the server. This is also one of the primary reasons our customers deploy Tableau. While the intention to house a single source of truth is there, we see deployments with a proliferation of data sources. This can occur when Creators are given full access to publish data sources at will, without any governance procedures in place. This results in a vast library of data sources that are often difficult to understand why and when they should be used.
The point is, Tableau Server allows for the creation of a data source that is general enough to cover multiple questions and flexible enough to address each of them without having to develop another data source. That being said, there are still instances when customers will continue to publish data sources on their server to fit the needs of the moment when. That data source may very well be a great stepping stone to improving the organization, if only it could be shared along with the details of what that data can help answer.
The other issue we see is data sources with odd naming conventions: something like, “ORAC_HIST_Sum_20”. There is someone, somewhere who knows exactly what that data sources holds and exactly what questions it answers. For the rest of us, it means nothing and we head to email or slack to find our next data.
Tableau Server Data Source Improvements
If any of the past sections apply to you or the server you are working with, do not worry. This is a great opportunity to shine up the data source repository and increase data source adoption for everyone. Here are a few things we have seen other customers do to make data sources more usable to end users.
Tableau recommends having a Data Steward or team charged with publishing data for the community that will meet their data requirements. There should also be a Site Administrator who will manage data sources, extracts and permissions when data sources are published. Once you have established a publishing arm and governance control move through the following steps to increase transparency of data.
Start with a clear name, something like, “Admissions Data Source”. Provide end users more information about this data source, which is where the data source details come in. Feel free to use this template for the About section of your data source:
Intended Audience: University Administration, Institutional Research, Finance, but the entire university can use this as a source of truth on admissions.
Business Use: The Admissions Data Source is the one-stop shop for admission trend analysis. Not only does it include information on college and degree program, but adds the granularity required to do class/faculty level analysis.
Connection Type: Extracted and fully refreshed every ~20 minutes.
Database: This comes from whatever database your school uses, specifically fact and dim tables. The exact schema can be seen here: https://yourinternalwebportal
What is a row of data? A row of data is a student, NOT a class. A class can consist of multiple student IDs.
Owner: Someone who is getting a raise
Backup Owner: Someone who was promoted
More Info:
Address the Project, Tags, and Certification
The project the data source lives in should be validated. Is this where it is supposed to be? Will moving it make this data more accessible? Will there be a need to change user permissions to a different project? We commonly see data sources get moved or promoted from one project to another once proper validation and vetting has taken place. This is all a process your internal team should be in agreement with. Once the data source is in the correct project we move onto tags.
Tagging content is a powerful way to help open up the search ease-of-use for your users and to take the pressure off of trying to create the perfect name for your data source. Any of these tags can be used in the quick search bar located at the top of your user interface in Tableau Server to locate this data source. Liberally add tags that will help your users track down data, so they can read and understand the well documented ‘About’ section you just created.
We will wrap up with the last, but certainly not least, data source certification feature. Certified data sources are special because they are the trusted pillars of data within your institution. Users who see this data source will notice a green check mark next to the data source name, indicating it has been certified. If users are using the quick search within Tableau Server, they will see certified data sources rise to the top of the list. If users engage with the filtered search, they can opt to easily view only data sources that have been certified, making their understanding of data in Tableau Server that much better.
Tip: If you would like to have end users see the information provided in the ‘About’ section above when they hover over the data source name or the certified icon, simply add those to the ‘Notes’ Section in the certification status:
This is what your end users will see:
Save Time and Money with Better Documented Data
The benefits that come from a clean data source structure are many. Survey users to find out how long they spend searching for data, what kinds of questions are they trying to answer, and who is the data source owner they reach out to in case of questions. Work with your Tableau Admin to look at the Server Admin Views within Tableau. Identify how often your data source is being used either through Tableau Desktop or via web authoring. When our customers are able to clearly label and educate end users on what data sources exist and why, they begin to see the impact of democratizing data.
Brian Russell
Brian Russell has been a Solution Engineer at Tableau for a little over 4 years. Brian first encountered Tableau while working as a Financial Analyst who was buried in the world of spreadsheets, v-lookups, and pivot tables (which he blames for the early graying of his hair). Utilizing a trial of Tableau Desktop, Brian was able to cut lead times for report building significantly and present new business insights to his executive leadership. He’s been a part of the Tableau team ever since, working with organizations in Higher Ed, K-12, and the commercial sector to help people see and understand their data like never before! *Opinions are my own and not the views of my employer*
