Shocker: ease of use requires expanding, not restricting, choices
Announcing a new venture

Making people jump over hoops

Take a look at the following chart, and guess what message the designer wants to convey:


This chart accompanied an article in the Wall Street Journal about Wells Fargo losing brokers due to the fake account scandal, and using bonuses to lure them back. Like you, my first response to the chart was that little has changed from 2015 to 2017.

It is a bit mysterious the intention of the whitespace inserted to split the four columns into two pairs. It's not obvious that UBS and Merrill are different from Wells Fargo and Morgan Stanley. This device might have been used to overcome the difficulty of reading four columns side by side.

The additional challenge of this dataset is the outlier values for UBS, which elongates the range of the vertical axis, squeezing together the values of the other three banks.

In this first alternative version, I play around with irregular gridlines.


Grouped column charts are not great at conveying changes over time, as they cause our eyes to literally jump over hoops. In the second version, I use a bumps chart to compactly highlight the trends. I also zoom in on the quarterly growth rates.


The rounded interpolation removes the sharp angles from the typical bumps chart (aka slopegraph) but it does add patterns that might not be there. This type of interpolation however respects the values at the "knots" (here, the quarterly values) while a smoother may move those points. On balance, I like this treatment.


PS. [6/2/2017] Given the commentary below, I am including the straight version of the chart, so you can compare. The straight-line version is more precise. One aspect of this chart form I dislike is the sharp angles. When there are more lines, it gets very entangled.



Feed You can follow this conversation by subscribing to the comment feed for this post.

Rhonda Drake

Wow, so much clearer what is happening after you rework it with the line charts! Way to go Kaiser!


I find the interpolation unnecessarily confusing. It makes it seem like there's a lot more going on then there actually is, and I keep trying to figure out what I'm missing.

The change over time is obviously much more clear in the % change version, but I still don't get a clear sense that the data is telling me anything important.

Perhaps the article deals with issues that make the data relevant, but my only reaction to the chart on its own is... ¯\_(ツ)_/¯ meh.

Dan Vargo

@jlbriggs You don't think the orange WF line dropping off a cliff is instructive given the context ("Wells Fargo losing brokers due to the fake account scandal, and using bonuses to lure them back")?


jlbriggs/DV: Thanks for the discussion. I brought that up knowing that some of you might object to it. The 3% drop in that period is the focus of the article and you can see why because this data tend to be very stable. What is missing though is any data about bonuses.


Dan - given that the UBS numbers could be said to have also fallen off of that cliff, and given that all four lines dip for 2017, no, I don't see the one line dropping as a specific standout.

I am not saying this is a problem with the chart itself.

I am saying that in simply looking at the chart I don't see anything that stands out as "yes, that is the message".

And yeah...I don't see how the number of brokers dropping tells us anything about how bonuses are being used to lure them back.


I hope that Kaiser eventually brings out a book showing the options for these graphs. Add in some exercises and some background material and it would make a great book for a course in business graphics.

I like the final graph but do wonder if straight lines would be better. Also would change in number of brokers be better, although probably not?

Andrew Gelman


Why do you call this a "bumps chart"? I'd call it a line plot or, more specifically, a time series plot. Am I missing something?


Andrew: I have an infatuation with these things called "bumps charts" that came out of the Oxbridge Boat Race: wrote several posts about it way back when the blog started (see here, and here). These have now been branded "slopegraphs" I believe by Tufte. But at heart, they are line charts with time on the horizontal axis, with a focus on changes over discretized time steps. Most often, the metric on the vertical axis is also discretized, most frequently ranks. "Bumps chart" is really a specific application of such a chart form to the "bumps" event at Oxbridge but to me, it's such a great example of the power of the chart form that I adopted that name.

Some blog readers have in the past complained that I use the term "bumps chart" when the data aren't discrete ranks but numeric but I see that distinction as immaterial. Also, this version here with the splines is not canonical as most such charts use straight lines.

Andrew Gelman


Hmmm . . . I don't see how this is different from what we usually call a time series plot.

I also agree with one of your other commenters that the curved lines are confusing. I prefer a graph that's more clear on what are the actual data points.


Added the straight-line version of it to the post so they can be compared.


"Some blog readers have in the past complained that I use the term "bumps chart" when the data aren't discrete ranks but numeric but I see that distinction as immaterial"

I'm one of those readers :)

I can't see how you could say that the difference is immaterial - it's the entire premise of a bumps chart.

I agree completely with Andrew here that this is just a standard time series line chart, and am actually rather confused why you would consider it anything else. Or, for that matter, how it is that you consider that to be synonymous with the very specific purpose and method of a bumps chart.

An another note, a slope graph to me is something entirely separate and distinct from either of those, and consists of only two data points per data series. A slope graph can certainly deal with ranks, but are by no means restricted to them.

Back to the graph at hand - despite the somewhat awkward interaction of some of the lines at the start of the graph, I greatly prefer the straight lines. It is immediately clear where the data points are, and what the patterns are.

The splines alter the data in such a way that it seems like there are many more data points that make up each series, and show fluctuations that aren't there. For data sets with many data points I sometimes use the spline instead, but here it seems a purely aesthetic choice that sacrifices understanding.


The comments to this entry are closed.