The only certainties are death death and taxes.  We might want to add to that the importance of data.  Since I own a business intelligence company, it's always interesting to watch as a group of people become interested in a certain subset of data, because inevitably they will come up with an ideal way they want to look at it. 

Traditional Visualization

Typically the data starts out in a tabular format that we're all used to from Excel and BI packages:

Once the data is available in tabular form, people start to identify the subset of data is that they're REALLY interested in, and dream up ways to visually represent that data that will tell them at a moment's glance the information they're looking for.  One way is to extend the tabular display via grouping, drilldowns, thresholds, and all the other ways that tabular data is traditionally enhanced.

Another typical way to do this is via a chart:

However even that presents too much data sometimes.  Usually after using charts for a while, people can pinpoint the subset of the subset that really has a maximum impact on their job function.  Rather than looking at historical trends, they want to zero in on that data and keep a constant eye on it.  To do this they typically set up something like a gauge so they can tell immediately if that key metric is where it's supposed to be:

That's about the extent of typically business intelligence visualization.  However, I've noticed some novel new ways to visualize data popping up lately, which is what this post is about.

Kicking it Up a Notch

Looking at large amounts of data and deriving meaning from it is always tricky, and people have come up with many different ways to attempt this.  The Web has opened up a whole new world of possibilities in this area, and new experiments in data visualization are popping up almost daily.  Some experiments are commercial, others experimental, others academic, others forms of art.

For information that can be represented geographically, mapping has become the flavor of the day.  This is primarily because it's so easy to build maps of data via mashups these days, but it also does a nice job of showing where things are happening, like this map from Zillow that shows homes for sale:

One kind of geographic data representation that hasn't seen widespread adoption yet (simply because it's much harder to do) is the technique of overlaying data on the map so that you can view the underlying data by looking at the map itself:

People have even started combining maps with traditional data visualization and chart methods, such as overlaying heat maps on Google Maps.

New Visualization Techniques

Some interesting new data visualization techniques have started being used recently.  One of the more popular and familiar ones is the "tag cloud", like the one I use on my site to show my categories:

Rich client technology has opened the door for more interactive ways to visualize data and watch it change in realtime.  Digg has been one of the more innovative companies in this area with its Digg Labs projects.  They created two brand new (to my knowledge) ways to visualize data, called Stack and Swarm.  Stack lets you watch the votes roll in in realtime to see the hottest stories:

The Digg Swarm view is also very cool, it allows you to see who's voting for which stories and watch the stories accumulate votes and grow larger with each vote:

I love how Digg is harnessing all the data flowing in to create interesting and fun new ways to keep people interested in what's happening on the site in realtime.

Another intersting use of Flash to present data is the We Feel Fine site, which uses to textual analysis to monitor blogs for sentiment and breaks down the data by gender, location, etc:

The color of the dots represent the sentiment that's being expressed, the size of the dot represents the size of the sentence.  Clicking on the dot lets you read the sentence that was analyzed.  We Feel Fine also has a few other innovative ways to present data, such as its Montage view.  The Montage view shows a montage of photos from people feeling a certain way at a given time.  For example this montage shows pictures from people who are feeling pretty:

Clicking on any of the images lets you zoom in and look at it in full-screen mode.  There are several other very intersting visualizations that these guys have created, I'd highly recommend you poke around because it's very fun and interesting.  One other thing that I absolutely love is the way they let you filter data.  To refine the population that's shown at any one time you use a highly customized filter screen that shows visually what you're looking at.  So aside from just seeing the data visually, they're also presenting a visual way to look at the filter that's being applied, which is a first as far as I'm aware:

Another interesting use of Flash to visualize data is, sadly, no longer available, but you can still read about it at the In The News page.  It was a tool that you could use to watch del.icio.us links flow in by category and get a feel for what was popular on any given day using an interesting "chunk" display:

Finally, there are some cool things being done with Flash by combining text and visual data in one visualization.  One such site is the newsmap, which takes Google News data and puts it on a treemap to display underlying patterns in the data:

So what's next?  I'm not convinced that all of these data visualization techniques will stand the test of time, but they're sure fun to play with.  And I'm sure some of the concepts driving them are useful and will be fleshed out further as rich clients gain more access to the wealth of data out on the Web.  I love seeing new applications of data and rich Web client technology, I can't wait to see how this space evolves over time.  New ways to look at data like these sites provide breathes new life into the data and makes it interesting again, and makes it more likely that we'll spot trends that we might have overlooked without them.

If you have any links to creative use of data visualization on the Web I'd love to see them.  Me?  I'm still waiting for Minority Report-style interactive 3D data manipulation, but I guess we'll have to wait for the new holographic displays to drop in price first.

I'll leave you with some links to a few other fun visualizations that I've found while writing this post:

• Vholoce:  Visualizes weather RSS feeds
• Synapsis:  Watch traffic flow between nodes on a network
• Musicovery:  An interactive jukebox that maps relationships between songs
• Moodjam:  An application that lets you track your mood and visualize it historically