By Ellie Fields 06 mai, 2009

We already wrote about the “Hey Look at This” moment that happens with Active Dashboards in Tableau 5.0. And there’s more in this release for visualization authors who want to create great experiences for viewers.

Here's a preview of some of the new authoring features, with the live interactivity that we're testing as part of Tableau Server:

Create Dual Axis Charts

Sometimes nothing but a dual axis chart will do. This is often true when you're looking at time-series data with two measures that are of different units and scale. In Tableau Version 5.0, you can satisfy your dual-axis craving with beautiful tableau charts.

Edit Legends, Titles and Tooltips

Customers have asked us for more control over all the elements of a graph. So we gave it to you. In the chart above, notice how the legend titles are customized. You can also choose to include or exclude data from tooltips.

More New Features for Rich Authoring

  • Create guided analytic workflows: guide your viewers through a set of links and filters to present the most relevant data depending on their actions.
  • Manage your data: use powerful new filters with relative dates and more styles. You'll see this throughout this and other blog entries on our 5.0 release.
  • Customize views: change Quick Filter and legend titles, customize tooltips and add parameters to titles and captions to present exactly the right view. We'll preview this today as well.
  • Create dual axes charts: show related data on dual axes. We'll preview this below.

We built Tableau version 5.0 to be a wonderful gift to anyone who authors visualizations. These are only a few of the more than 60 features that are coming out. Stay with us!

Next week: How big can a Tableau implementation get? Much bigger than before. Tune in for server scalability.


When using a dual-axis, is there a way to bring the highlighted marks/lines to the front? If not, I would recommend adding some transparency so that when the marks do overlap, the one below are still somewhat visible.

Also, Stephen Few recently went into detail on Dual-Scaled Axes in Graphs looking for an example where they would be a valid comparison, and could not find one.

Your example use of a dual-axis also generates invalid comparison as Stephen Few points out. I believe a single axis with a percent difference from first would generate a valid comparison (Tableau makes it very easy to do this). Your dual-axis seems correct for the previous 5 years, but the invalid comparison becomes more obvious with a longer range (see attached 5/15/25 year range comparisons).

What do you think, is this a valid comparison, or when is a dual-axis a correct representation of the data?

When charting in dual axis, is there any way that Tableau can mix two different chart types (ie. bars and lines)?

It's amazing. I'm quite excited by this next release. :-)

THe "Select Date Range" filter is pretty neat - teach me?

The dual axis and highlighting great, and the in-situ parameter selection is very nice. Two small suggestions:

1. Keep the frame. When the selected data values are out of range, the graph frame disappears (I tried 2 days). This could be disconcerting. Instead, it might be good to show the frame, without labled axes, and add a "data out of range" message.

2. Toggle, not just gray-out. The mode where the non-highlighted series are grayed out is very nice,since it helps point attention to the data of interest. One idea we found worked very well is to toggle the data series in and out, so that non-highlighted data don't interfere, but also so that various comparisons can be made by dynamically flipping them in and out.

Thanks for all the comments and suggestions.

@Joe: Good comments, and you're right, dual axes have been a source of great debate in the visualization community. We finally decided to offer them because in some situations, especially with time data, we do believe they can be useful: for example, here you can see the the Dow and the Nasdaq had similar peak levels over the last 30 years, but about 8 years apart.

@mfan: No, at this time we don't support mixing bars and lines.

@Meliza: The new relative data filter isn't hard- when you get a copy of Tableau 5.0, you'll see it as a Quick Filter option just like any other Quick Filter. If you have trouble finding it, just give us a call.

@bernicerogowitz: great suggestions. They're both on the list of future enhancements for this feature, so we're thinking the same way you are.


Ellie Fields,

Thank you for your reply. I would like to reiterate one of Stephens's findings (linked to above):

"It is inappropriate to use more than one quantitative scale on a single axis, because, to some degree, this encourages people to compare magnitudes of values between them, but this is meaningless."

When you say that with the dual-axis chart "you can see the the Dow and the Nasdaq had similar peak levels over the last 30 years, but about 8 years apart." I would contend that statement is not true. The set of peaks in 2000 and 2007 are very different, and a dual-axis of this data set leads the viewer to incorrect conclusions. I agree that in 2007, the indexes moved in a very similar fashion, but around the year 2000, Nasdaq spiked in a much greater degree and eventually returned to a similar trend. Also with the "Single Axis - % Diff from First" the viewer can see that Nasdaq also dipped relatively lower around 2003 (see the above attached 25_Years.png).