One of our clients recently brought us on to help with a BI project that had been started by another consulting firm. While the competing firm did some great work, they just were not able to keep up with the demand for quick iterations, and expansion of analytics and Business Intelligence within this organization. They put together some great looking dashboards, that were pixel perfect versions of some wireframes that they had put together over a year ago. This was all done in Tableau.
I was first introduced to Tableau in a project I did for Skype way back before Microsoft acquired them. I was truly amazed with Tableau at the time and was able to quickly iterate over requirements and push out executive dashboards in a timely fashion. Now, I’ll have to admit, these all used Out Of The Box (OOTB) visualizations and esthetics of these dashboards were not first and foremost. Tableau provided an easy way to build these dashboards with minimal fuss.
Times have changed however. The modern consumer of analytic dashboards wants customized and branded dashboards along with quick iterations. Unfortunately, Tableau seems to have gotten a little “long in the tooth” when it comes to their viz libraries. Tableau is 100% customizable, but building things as basic as a gauge chart to see actuals VS a goal takes a ridiculous amount of time. I don’t want to have to bill hundreds of dollars an hour figuring out what the sin/cosine of an angle should be to visualize different colors and goals on a gauge chart! It’s not intuitive, it’s not fast, and it’s uber frustrating even for someone that has used it off and on for over a decade.
I have also had the opportunity to use Apache Superset, Amazon Quicksight, and Looker on a few different projects in the past year. Amazon Quicksight seems to be super stable, updates come out nearly every week, and the user community seems to be growing quite well. Amazon’s “Tableau Killer” seems to be getting some traction!
About the only thing I have found that doesn’t work well in Quicksight is joining a snowflake data model. A star schema is easy going from a fact to a dimension, but if you want to go beyond that to a sub-dimension, you have to use multiple datasets to do that. That lends to confusion and dataset bloat.
I still believe that Tableau is a great tool, but take a spin with Quicksight and see how much more productive you can be!
Want to find out more about getting value out of your data? Click “Get Started” and let’s go!
Comments