Why I Still Show Up to Tableau User Groups After 19 Years

I've been using Tableau for 19 years. And I still show up to Tableau User Groups.

Not because I have something to prove. I show up because user groups are one of the best places to hear what experienced analysts are actually building, pick up techniques you wouldn't have thought to search for, and connect with people working through the same kinds of problems you are. That's true whether you've been using Tableau for six months or sixteen years.

The talk I gave at the Analytics TUG in May 2026 focused on customer retention visualizations, specifically how to build jump plots in Tableau. A jump plot uses semicircles to show each customer's lifespan, with the height of the arc scaled by a value like total sales. Profitable customers arc upward, unprofitable ones flip below the axis.

It's not the kind of chart you find in Tableau's default chart picker. You have to build it deliberately, from the geometry up. And that's exactly what makes it worth sharing.

5 things worth watching for:

1. Save your geometry calculations in Notepad and reuse them forever (4:42)
The jump plot is built on two formulas: radius x cosine(angle) for columns, radius x sine(angle) for rows. Write them down once, save the file, and copy-paste them into every future workbook. No need to rederive them.

2. Use a relationship, not a join, for the points data source (9:20)
The second data source is a single column running from 0 to 180, one row per degree. Connecting it via a relationship means Tableau only pulls in those 181 rows when needed, rather than padding your main data source 180 times the way a join would.

3. Convert your dates to integers before you start building (15:00)
Tableau cannot draw jump plots using date fields directly. You need to wrap your start and end dates in INT() and use those calculated fields instead. Make both of them dimensions to stop Tableau from trying to aggregate them.

4. Use SIGN() to flip unprofitable customers below the axis (28:09)
Wrapping the profit calculation in SIGN() returns 1 for profitable customers and -1 for unprofitable ones. Multiply that into the rows calculation and the unprofitable arcs flip downward, making them visually distinct without needing a separate sheet.

5. Use MOD to cut mark count when performance is slow (46:35)
If your view has too many marks, filter the points data source with [Point ID] MOD 2 = 0 to draw only every other degree. The arc still looks smooth and the mark count drops by half. You can push it to MOD 5 if you need to go further.

Every time I build something like this in front of an audience I notice that the moment people see that it's possible, something shifts. They stop thinking about Tableau as a tool for making the charts they already know, and start thinking about it as something you can use to answer almost any question if you understand what's happening underneath.

That's the difference between analysts who plateau and analysts who keep getting better. It's not talent. It's whether you're still building things that push you.

A viz like this is worth adding to your Tableau Public profile. Not because it's complicated, but because building it properly means working through the data model, understanding how calculations behave at different levels, and solving a real business question from scratch. That kind of practice compounds over time. So does showing your work.

If you're looking for more of this kind of thing, that's what Next-Level Tableau is built around.