Applied Zen in Tableau (Conditional Formatting)


Published on

Presented the Boston Tableau Users Group Meeting on May 17, 2012. The workbook that goes along with this presentation is on Tableau Public at If you have other ideas, questions, corrections, or issues, I’m glad to respond, email me at qm.jtd (at) or on the Tableau forums at

1 Comment
No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • This was the presentation I gave at the Boston Tableau Users Group Meeting on May 17, 2012. The visualization that goes with the presentation is available at a dozen years ago I took a break from working in high tech, lived in retreat centers and intensively studied Buddhism. A fundamental concept in Buddhism is that the duality we see is really an illusion, for example…
  • We’re often stuck on a problem where we only see two choices, and both of them suck. This way doesn’t work, and that way is really no way to go.Zen Buddhist stories are full of students walking up to their teachers, saying “Master, I’m stuck in some duality”, and then the master whacks them upside the head and the student is enlightened. The master isn’t stuck in the duality, and is free to take radical action So, for us…
  • There’s always another way. You just need to whack your brain, your data, and/or your viz with a stick.
  • Most of the people I work with are used to laying out data with Excel. Excel starts with the presentation, you choose bold, background color, etc. and paint your formatting onto the screen. It really munges data and presentation together.
  • Tableau starts with your data, and from attributes of that, defines the visual elements and their meaning from attributes of the data. So a bar isn’t just red to make it red, its red because it’s marking the lowest performing salesperson.So, a number of formatting functions that are very easy in Excel take more work in Tableau, because we need to set up the data. And some are just more difficult because the Tableau developers have been working on other things.Now we’re going to switch over to Tableau. Anyone on 6.1?The workbook I’m showing you works in both 6.1 and 7.0, with one exception that I will mention. If you are online, you can download it from:
  • Given that Tableau starts drawing from the data, we need to think about the data in a different way to get to the presentation that we want. The way I’ve internalized this kind of thinking is to call it “orthogonal” thinking and what I encourage you to do is think beyond the normal ways of drawing marks on the screen. Tonight I’ll be showing you some techniques for working with text tables and labels. I won’t have time to go through all the details of how, the instructions are posted in a workbook on Tableau Public at here on 6.1?
  • How I got here…Last summer I started as a Data Analyst in QM at SMMC150 bed hospital and 18 physician offices in southern MaineFirst data analyst in a department of 14, pretty much all their data in Excel spreadsheets or on paper, getting overwhelmed with requestsIS dept. super-busy keeping the duct tape replenished on existing systems, implementing shared electronic health record and new HR systemMy job has been to organize the data, teach staff how to work with the data, and find ways to share the data.So I’ve been busy.And, I’ve been lucky that this is a new position so I’ve been able to help write my own job description. Part of that was identifying solutions and I created the analysis and business case for buying Tableau.Using Excel, Access, started using Tableau in September with 1 Desktop and 1 Server licenseI took on mastering Tableau as a personal goal, and part of my learning process has been to use the forums as a source of problem sets, I’ll talk more about that process at the end.~1500 metrics and measures that we trackOnly a few have a direct impact on revenueMore metrics coming, more ties to finances (Affordable Care Act of 2010), more comfort with data in the organizationWe’re big enough to have some really interesting problems, small enough to have a real impact.
  • Which doctor is the best? This sample data set is pulling in metrics from 5 sources, including patient care information and patient surveys. The chart at the bottom with the green shading shows spread, where there’s a large spread for those of you who do Lean or Six Sigma will know that we don’t yet have standard work in place. We’re using this information to stimulate friendly competition among the providers to do their best and to highlight areas for attention.Pushing the envelope on table calcs and partitioning
  • And a dashboard for infection prevention that we are just starting to roll out on Tableau Server. Used to be updated quarterly, is now real time, and adds bells and whistles like Days since last infection.Makes for some interesting views because not all data elements needed for the calculations are there, so we have to do assorted padding of the data to get the numbers to come out right.
  • Readmissions are difficult for both the patient and the hospitalHospital has a three year goal of reducing readmissions by 10%Moving from 6+months out of date #s to near real-time tracking, this has been huge for us to shift from a rearward view to a projection.Loaded a couple of hundred dimensions and looking for patterns in our data (by making lots of views), here’s one that has jumped out. Where the day of discharge for the patient has a relationship with their chance of being readmitted. It turns out that this has a lot to do with what kind of care patients get after they leave the hospital, so we’re working on those transitions.
  • Applied Zen in Tableau (Conditional Formatting)

    1. 1. Applied Zen in Tableau
    2. 2. This No wayway no other optiondoesn’t really dowhat I want
    3. 3. This No wayway Another way Whack your brain, your data, and/or your viz with a stick
    4. 4. Excel starts with presentation“Paint” formatting onto the screen
    5. 5. Excel starts with presentation “Paint” formatting onto the screen Tableau starts with dataVisual elements (and their meaning) derived from attributes of data
    6. 6. Thinking Orthogonally• Going beyond the “usual” ways of displaying your data – Mark labels to show numbers and text – Dual axes to show calculated trend lines and reference lines – Custom SQL to pad and reshape on the fly – Table calculations for fun and amazement• There’s always Excel / OpenOffice
    7. 7. Southern Maine Medical Center150 bed hospital and 18 medical Quality Management - tracking >1500offices located south of Portland metricsBig bang implementation of new Using legacy systems, paper, SaaS,medical record, patient care and HR Excel, Access, and Tableausystems in 2013
    8. 8. ProviderRankings(sample data)
    9. 9. InfectionPrevention(sample data)
    10. 10. ReadmissionsAnalysis(sample data)
    11. 11. Conditional Formatting via the Multiple Axis Crosstab See the viz at gv4/Introduction
    12. 12. Multiple Axis Crosstab• Instead of using the Text mark with no axes on the Columns shelf and Measure Names as the way to get multiple columns, we have a “fake” axis for each column and use Text, Shape/Circle/Square and/or Gantt Bar/Bar marks and mark labels to show the values.• Advantages over Measure Names/Measure Values: – Can mix text and numbers, left and right alignment – Each column/row can be formatted differently using Text, Shape, or Bar marks – Can even use Bold, Italics, and Underline for calling out text• Disadvantages – Extra work to set up – Extra white space in the view – Harder to see whats going on in the view, since Measures are on the Label shelves and not on Rows/Columns or visible on the Measure Values Card – Sensitive to display size (particularly for Shape/Circle/Square marks) – Slower performance due to making Tableau do more work to display data
    13. 13. Mastering Tableau• Pick an area – formatting, mapping, table calculations, survey analysis, data reshaping, dashboards, Tableau Server, big data, etc.• Use Tableau training resources – Training videos, do the online training, read the knowledge base articles – RTM, Rapid Graphs in Tableau book• Practice, practice, practice – Tableau forums are a great place
    14. 14. Links• Online Training -• Tableau Forums -• Rapid Graphs in Tableau -• Tableau Blogs – – – – – – – – – – – – –• Data visualization: Edward Tufte, Stephen Few, Naomi Robbins, Nathan Yau, etc.
    15. 15. The Other Kind of Visualization
    16. 16. Comments, Questions? Jonathan Drummey @jonathandrummey