Posted by John Strykowski
March 2, 2017
Creating Power BI dashboards can seem easy at first, but the more you learn the more you appreciate that with so many options and choices, there are many complex decisions to be made. And very specific skills are required to go beyond simple visualizations. This article summarizes the top four skills you’ll need to excel at dashboard design and development.
You will need a really sound understanding of the base theory behind data visualization and dashboard design, preferably in a software agnostic context.
I have been designing and implementing dashboards professionally for the past few years, and there are two books that have helped me become the “dashboarder” I am today:
We’ve developed UD3, our own methodology for Dashboard Design and Development. This process scrupulously follows the industry’s best practices for dashboard design and implementation. UD3 ensures our dashboards not only look great but, more importantly, enable our clients to uncover critical and actionable insights.
Having experience with any relational database and SQL definitely ranks high on the list of recommended skills.
Basically, what I’m hoping to convey here is the importance of having some experience with Structured Query Language (SQL), be it with Microsoft Access, MS SQL, MySQL or even Oracle. The inter-web abounds with resources that will allow you to quickly get up to speed. When working in Tableau, it’s important to have at least a passive familiarity with the different table joins. When people talk about data types, you should know what that means. If someone hands you some data and tells you that it’s in tabular format so you’ll probably need to unpivot it, you want to be able to say “no problem” instead of passing out from stress.
Two careers ago when I was a web developer, PHP and MySQL were pretty much considered arcane knowledge by Joe Public. These days most marketing professionals are familiar with at least one Content Management System (CMS) and have basic knowledge of HTML and CSS. With tools such as Power BI, it becomes easy to create simple visualizations that would have required an engineering degree a few years ago. With that in mind, I’m going to assume that getting at least a basic understanding of how data works in a structured environment will soon become much more common and even mundane in a few years.
As a web developer and because more and more web sites were becoming database driven (yeah, it was a while ago), I decided to teach myself how to write a store procedure and how to query a database to generate content on a web page. What I managed to pick up in my previous (professional) lives has been invaluable. Give yourself the means to understand data even if only at a basic level. It won’t let you down; data’s your friend.
Data comes in all shapes and structures, and it’s up to us to make it Power BI-friendly. ETL stands for Extract, Transform, and Load. We do love our acronyms! This is what might happen to your data before it gets into Tableau. Why? Because more times than not, the data you’ll be given isn’t in good enough shape to simply be imported into Power BI. Now that being said, ETL is vast and complex, and I’m certainly not suggesting it’s critical you devote the next two years to acquiring those skills. But it’s important to understand and recognize that data comes in all shapes and forms, and that sometimes the data will need some massaging before it’s ready for Power BI-consumption.
My background in ETL comes from the data mining I performed on digital activity logs when I was working in Digital Analytics and had to analyze visitor traffic in log files to discover meaningful patterns or the odd activity relating to some not-too-kosher website activity behavior. I would take a 2GB text file and upload it into MS SQL and start exploring the data. In other words, I had the “loading” part down, but that’s about it.
The ETL learning curve was a little steeper for me. I’ve been working with Alteryx for the last little while and finding it to be an intuitive and full-featured application. Talend is also a good option, especially since the basic version is open-source and free.
I didn’t exactly have a lot of experience in these fields prior to my debut in dashboard design and development, but I did have a quantitative background in that I’ve almost always had jobs that involved some degree of math and knowledge of computational techniques. There’s no secret formula here, it comes from experience accumulated throughout my career.
I happen to be in an ideal situation where my department is filled with people of various backgrounds with similar base skillsets when it comes to dashboard design but with a wide variety of specialties. We collaborate and constantly learn from each other. The whole is stronger than the parts, etcetera, etcetera.
That being said, I can recommend a couple of books:
I’m a fan of the Head First series. When learning by myself, I tend to retain information more easily with books of that collection and across a wide variety of topics, so they must be doing something right.
My last reco, before you go out and spend your dollars on books and training, is to check out the UD3 methodology page on our site. It’s a quick read and most people find it quite informative. Please don’t hesitate to contact us should you have any follow-up questions.