Thursday, April 2, 2015

Line charts aren't just for time series data

The tweet below is getting a lot of attention today. It's true, as Katie and others suggest in the discussion, that the chart can be better why ordering the music genre that way, to begin with? but line charts aren't just for time series data. This is a very common misconception; I still remember my Math middle school teacher telling me exactly that, back in the Pleistocene (ahem)!

Time series charts are just one type of line chart. There are others. Parallel coordinate charts (read 1, 2, 3, 4, 5) can be tremendously effective to reveal patterns.

(The graphic on the cover of The Functional Art, by the way, could be perhaps considered a parallel coordinate chart with just two vertical axes...)

1 comment:

  1. As you pointed out Alberto, the first flaw regarding this chart is that the ordering of the "levels" along the X axis is totally arbitrary. Actually, if the point they wanted to highlight is the ordering in terms of life expectancy, they should have somehow ordered the levels on the Y axis! (the X dimension here is totally spurious!)
    But what strickes me most is the surimposition of the US life expectancy. This makes even less sense that the first (bold) lines -if this is possible... What on earth is the X value for these lines linked to? Musical genre? It certainly looks like a historical evolution of life expectancy, but on what basis do they surimpose the two?

    This being said, I do not agree with Katie's comment -line charts aren't only for time series.