Their primary goal was to celebrate Obama and to deliver a political message: the reinforcement of the “Obama is popular in the US and elsewhere” and “Obama is a uniter” narratives.

If one had focussed primarily on the goal of understanding the data better, rather than celebrating Obama, it probably wouldn’t have occurred to them to multiply the numbers to support the “one people” part of the narrative… but the fact that it was “not fair” as mentioned above, is secondary to making a pleasing visualization that delivers the intended message… even at the risk of misleadingly representing the underlying data.

They had different priorities. I don’t think one type of goal of a visualization is “right” or “wrong”… they each have their place. I hope at MIT they are learning the difference though… and not learning that it’s ok to start with a desired premise and then fit the data to… and then call it “science”. Perhaps that’s part of why we expect scientific rigor… simply because it came from MIT. Had this visualization been produced from some liberal think tank, would we have commented in the same way? Or would we have just enjoyed the visualization for it’s own sake?

If you want to criticize how well this visualization works from a scientific point of view, it seems there are bigger problems than “non-US traffic is overstated”.

I think the main thing here is that most people are going to watch the video and think it’s the same scale, because there’s nothing in the actual viz that suggests otherwise.

If the primary use case was to show rates of change (versus total volume) and you have one group of data that outlies other data then perhaps you use a log-scale versus linear…. is that violating the integrity of the visualization even if it distorts the data?

So to your case above, if the goal is to do a comparative analysis between country data, then no, don’t scale the other countries. But, if the goal was to do a comparative analysis of rate of change across countries, and then a finer level comparative analysis intra-countries then perhaps the approach Andrea took would be appropriate.

In those situations, where the “integrity” of the visualization may be questioned and/or distorted by the scales being used and it is an interactive visualization I often give users an option to toggle between scale modes so they can see the data both ways.

I think it’s also a bit pretentious to replace “cell phone traffic” with “emotional flow”. I might as well call Death and Taxes a Visual Guide to Presidential Desires.

I know the design choice was deliberate, which is why I felt compelled to highlight it.

I don’t think it’s very common to represent variables of different magnitude simply my multiplying one. If that is common practice, it shouldn’t be.

For simplicity’s sake, let’s pretend we’re dealing with a bar graph instead of a map. We have bars for the U.S. and then bars for foreign countries. In our original bar graph, we see that growth for foreign is really tiny compared to the U.S.. Do we increase the foreign bars by 10 to make it look better and then add a footnote? No.

It’s the same idea with the map. It’s simply not a fair representation of the data.

Have a good day.