<-- Advertise here.

Subway Sparklines [transit.frumin.net] illustrates the relative yearly ridership of every New York City subway station from 1905 until 2006. Each station is represented as a separate sparkline, located on the New York city map. The general idea it that the history of subway ridership tells a story about the history of a neighborhood that is much richer than the overall trend. An example shows the comeback of inner Williamsburg, and how the growth decays at each successive stop away from Manhattan on the L train.

For more information, check out the map author's accompanying blog post.




This is a gratuitously wrong use of sparklines.

The data vis is made for a web site, which means the sparklines here are created to be displayed on a (comparatively) low-resolution computer. They totally lose the quality of "visual confection" that sparklines were originally supposed to have. Furthermore, the haphazard positioning of the sparklines makes it impossible to compare multiple sparklines in any coherent way. Also, because they're scattered across the screen, there's no baseline for the sparklines -- they're floating in space. This isn't a problem with in-line sparklines since the text serves as a baseline nor with multiple sparklines stacked and aligned with each other but for this visualization they're just all wrong.

This vis would be much better served with an area graph that colors (translucently?) between the baseline and the top "sparkline" -- the size of the colored area would allow a quick idea of overall traffic over time, interesting data features would be more prominent and the end result would be much more readable.

Thu 14 May 2009 at 10:02 PM

While I disagree with some of Tuftian's comments above (I wonder: how would Tuftian define "the quality of 'visual confection'" and show the resolution at which this quality is lost?), I do agree with one point. The display would benefit from a tabular comparison of the sparklines. As is, the map display really only facilitates an analysis of the change in an individual stop's ridership over time. I don't know how one could say that this information isn't useful (individual riders may only be interested in looking at data from stops they use, and the map format makes it easy for them to find the data), but Tuftian is correct in noting that it doesn't allow for comparisons between stops.
However, it is a really intriguing application of sparklines; I'd like to see more projects like this.

Fri 15 May 2009 at 7:35 AM

The display can't benefit from tabular comparison, since that would defeat its original intent to make use of the positional relationship of the data on a map. I agree with Tuftian in having transclucent backgrounds for each sparkline that can both bring out the baselines and provide better contrast and separation from the underlying map.

Fri 15 May 2009 at 2:58 PM
Zen Ho
Commenting has been temporarily disabled.