The buzz

Jittering lines

A reader alerted me to this NYT chart a few weeks back.  The chart plots daily changes in stock index prices (gray lines) and yearly changes (color blocks). 


The blue blocks represent bad down years but notice that the daily changes during many of those periods give no such impression.  Nyt_volatility2_2In fact, the gray lines are quite equally balanced on both sides of 0, and yet the annual tallies swing from positive to negative quite frequently.  It is by no means true that one exceptional down day predicts a down year.

The problem arises from cramming too much data into too small a space.  We can't judge the density of the lines on paper and so can't judge whether there were more up lines than down lines.

This issue is not dissimilar to the jittering question when used with large data sets.

Source: "The Pulse of Uncertainty", New York Times, Jan 4 2008.



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

Jorge Camoes

Tufte likes it ("Brilliant news graphic by Amanda Cox")...


I think it's ok. I think you have sort of assumed that a bad year will co-relate with bads days. Looking at the chart I think that after a bad year the daily fluctuation increase. The density should be equal? - one line per day no?
It reminders me of a ECG readout plotted over tidal volume bar?


It should be possible to represent the density of lines that are too close together to distinguish, by using gray scales. The lines appear to be about one month in width, so you want about 30 levels of gray for maximum precision (~30 days in a month). If only a few gray levels are available, that would still be an improvement over black/not black.


When I saw the chart, I had no problems with it. My immediate interpretation was that they gray lines were a way of depicting volatility to an audience that may not have an as intimate understanding of the concept as readers of this blog.


It's not bad at all. That doesn't mean it couldn't be even better.

Re: my previous suggestion of gray scales, an alternative would be to abandon lines altogether and have dots instead. Thirty dots per month would show the distribution nicely.

But removing the lines would probably mask the monthly volatility: the eye would naturally connect the outlying dots with each other across months, instead of with the inlying dots in the same month. Also, the inlying dots would probably be dense enough to still mask each other from overcrowding.


It's not about whether this graph is good or bad. It points out a difficulty with density of lines, dots, etc. If the annual color blocks were not available, one would use the gray lines to form a judgement as to how the overall index would be doing. This would lead to mostly wrong inferences is the point I'm trying to make.

So it is to their credit that they provide both daily and annual data series, allowing us to observe the relationship between the two!

The comments to this entry are closed.