Fun with Graphs.

On the reccomendation of my friend Holly, I’ve been reading the excellent SixthW blog, which concerns itself mostly with the effects that the web is having on journalism – a topic in which I also have an interest.  I was clicking back through their recent history and I came across this entry, six or seven posts down, about the now-not-so-new (but still cool!) New York Times Visualization Lab.

What the Visualization Lab does, basically, is provides you with some data sets – say, the county-level 2008 election results, or the words in Obama’s inauguration speech – and it lets you graph that data in one of a dozen or so ways.  For example, you can graph the election results as a word cloud (much like the one on in the right-hand sidebar), where the size of the state names is relative to the total number of votes cast in that state.  It looks like this:

(Note: This graph is a little confusing because its colors look like traditional Democrat and Republican colors, but it really is just total votes cast.  So you can see that states like, say, California and Texas had a larger percentage of total votes cast than did Utah or Wyoming.)

I’m not saying this is the most useful graph ever.  And the service has other problems: the interface is clunky; there are far too few data sets; and the data sets that are there are somewhat useless.  But there are a couple of things that are still really cool about this, and the biggest one for me is that the New York Times is acknowledging that the quality of news can be improved by further interaction between news-provider and news-reader.  For example, before they implemented this service, they would still have had access to this information, and they probably still would have provided it to me with some sort of graph.  But they almost certainly would have gone with a more traditional bar graph or a pie chart, and the problem with those is that people process information very differently.  A chart like this is much easier for me to quickly (if generally) grasp than, say, a pie chart is.

The Visualization Lab is cool, but obviously it needs some work.  Which is why I would really recommend that you click through to the Many Eyes website, where the technology was originally developed.  Unlike the NYTimes, it lets you upload your own data sets – so you can create snazzy graphs of just about anything you please.  Check these out:

For this first graph, I counted the starting letters for all the artists in my iTunes library, and graphed that as a similar Word thing:

Obviously, “S” is rather large, but I was surprised that “B” came in such a close second.  Letters like “X” and “Q” are very, very small – you might actually mistake them for specks on your monitor.

Somewhat more usefully, this word tree illustrates what I was saying last week about Obama’s inauguration speech; I typed in the word “we” and it showed me all the things that followed it.

Expect to see a lot more arbitrary graphs on this blog in the future.  And if you click through to the graph links, you can click around all of them and zoom and see different data values and things.