Background: How Many Eyes Evolved

The idea for Many Eyes came from some surprising behaviors that we each serendipitously observed.

Several years ago, Fernanda was working on an email visualization program. When it came time to run experiments, she was extremely careful to let her subjects know that visualizations of their email would be completely private - the assumption was that no one would ever want to reveal their personal mail. To her surprise, her subjects immediately began to ask for ways to share the visualizations with others! In fact, it turned out that the process of storytelling and reminiscing was one of the most valuable aspects of the visualizations.

Martin's serendipitous experience came two years later, when he created a baby name visualization to illustrate a book that his wife had written. After the site went live, he spent an embarrassingly long time doing Google searches to find out what people were saying about it. His self-centered surfing was rewarded with the discovery of many large, detailed blog conversations in which users - who often had no immediate interest in naming babies - speculated about various trends and patterns they'd found in the data. In aggregate, the analysis of the data was both deep and broad - uncovering a huge amount of information.

These two experiences were the motivation behind Many Eyes. We wanted to find out whether these experiences were flukes, or whether there really was a powerful social angle to visualizations. And we felt that the only way to find out was to create a participatory website available to the entire internet - to create not social software, but societal-scale software.

Many Eyes Visualizations homepage

Social use of visualizations - what's happened in the last month

Many Eyes was launched a little more than a month ago. By the way, the original Read/Write Web article listed "Fernanda and Martin" as the developers - in fact everyone in the IBM Visual Communication Lab has worked on the site, and probably the two of us did less of the actual development than anyone else! 

So what has happened? Was our hypothesis about the social use of visualizations correct? It's too early for any scientific conclusions, but that won't stop us from giving some anecdotal evidence ;-) Here are three anecdotes:

1. One of the earliest users of the site uploaded a set of data that described which figures in the New Testament were mentioned together, and then used our graph visualization tool to create a network diagram of these figures. You can see this visualization here (screenshot below) and the blog post they wrote about it here. This network got picked up by a whole community of bloggers who were interested in the statistical analysis of the bible - we counted more than 100 trackbacks on the ESV blog where it was posted. 

New Testament figures visualization

This in turn led other users to upload their own data to create visualizations, such as the proportions of New Testament authorship:

New Testament authorship visualization

2. Another user uploaded a set of data on global warming. In this case, most of the discussion occurred on the site itself - where there was a heated (if you'll pardon the pun!) debate over the meaning of CO2 and temperature trends in recent years.

Global warming visualization

3. A frivolous example, that nonetheless shows the social aspects of the site, can be seen in this visualization of books. Entitled "Harry Potter is Freaking Popular", the diagram is a bubble chart that shows the top 50 books on LibraryThing. In this case, the comments hold not much analysis. But we started a little game: we realized that we could use the highlighting feature of the chart to show which books a person had (or hadn't) read. As of this writing, a dozen people had participated in the "game" of showing what they had or hadn't read.

Library books visualization

What's Next?

A lot, we hope! We think that social data analysis is a lively area right now and we are not the only ones exploring this space - two other sites of note are Swivel and Data360. Each of the 3 sites has a different emphasis, but what we have in common is a belief that the web enables a new, social kind of data analysis; a type of statistical thinking that is both playful and serious. We'd argue that this is just one way in which visualization is becoming a new and important mass medium - but that's a blog entry for another day...

Swivel

Data 360

The software, Glasshouse by Green Phosphor, lets you take data from either a spreadsheet or database query and place a 3D representation of it into a virtual world environment where it can then be explored interactively. Users are inserted into the virtual world as an avatar which can then manipulate the visualization of the data by drilling down into it, re-sorting it, or even just spinning it around to see it from all angles.

The benefits to working with data in this way don't really need to be touted too much - many businesses already perform data visualization, often using expensive software and powerful computers to do so. What makes what Green Phosphor does so interesting is not that they've come up with a way to visualize data - it's that they've come up with a way to leverage the platforms of virtual worlds to do so.

How it Works: CICP (Think HTTP for Virtual Worlds)

Some of the company's solutions involve using a proprietary virtual world, "Glasshouse," for data visualization, but for Second Life, Sun's Wonderland, and other virtual world users, they've developed adapters that project graphs from Glasshouse into whichever virtual world you're using. The only requirement is that the virtual world be CICP-enabled.

CICP, or Content Injection and Control Protocol, was developed in-house by Green Phosphor CEO Ben Linquist and released to the public domain. The standard, cross-platform protocol essentially serves as HTTP for virtual worlds where it works as a communication mechanism that the Glasshouse gateway can use to generate temporary artifacts in the worlds. Already it has been added to Sun Wonderland and released under the GPL license there. It has also been implemented in Second Life with the help of a Java servlet and released under a BSD license. The company is currently working to add it to other virtual worlds, too.

Data Viz for Anyone: From Spreadsheets to Biotech

Depending on company size, there are three different levels of service available. First, a spreadsheet world lets you upload Excel spreadsheets that can then be visualized in a web interface. Next, there's a workgroup appliance that delivers data visualization and virtual conferencing needs to small or medium-sized businesses. And finally, enterprise solutions designed especially for virtual markets like bio-technology have also been developed as more customized solutions.

As Linquist explains in this YouTube video, the technology is even advanced enough to produce a virtual laboratory where researchers can perform model-based drug development.

If you have Java installed, you can test their web-based virtual world demo by clicking here (launches Java window). For more information about their solutions, visit GreenPhosphor.com.

For World Oceans Day, Google has come to the rescue with a cool collection of visualization apps that use Google Earth to encourage analysis of the big blue. According to last night's post on the Google LatLong Blog, "To help get our heads around the watery part of our planet, we've collected a few visualizations of both issues facing the ocean and cutting edge science. Have a look at the problem of plastics in our oceans, find out about where currents might take your trash, the health of the world's fisheries, and what the latest satellite data says about the state of the ocean."

According to Google, this makes visualization apps even more important to our understanding and protection of the world's oceans. The related Google Apps page also shows a news ticker with World Oceans Day items and a custom search box in the upper right to find content from partner sites.

The apps show sea surface height, sea surface wind, areas of overfishing, oceanic currents, and other non-Google Earth projects that graphically illustrate marine pollution.

The Google team has also curated a collection of relevant videos on YouTube. While some of the whales/sharks/fish vids make for pretty interesting browsing for fans of shows such as Planet Earth, others serve as graphic and sad reminders of humans' impact on the oceans.

Using the Tulip visualization package, he sampled 25,000 relationships out of a data set of 166,332 users pulled from last.fm with help from their API. From this data, he came up with 2310 seed users with 19,008 friends, resulting in 24,036 relationships and an average of 10.41 friends.

From the resulting visual graph, it was clear that last.fm's user network is actually very strong. Although there are some clusters of close friends, even those people have a lot of interconnectivity:

Along the outside of the main network, he found users who either did not have connections within the main network or whose connections were missing because he was only looking at a sampling of users as opposed to the full network. Likely, it was a combination of both. Interestingly enough, even among these outlying users, they had smaller networks of their own:

Graphing the distribution of users to number of friends, it was clear that the majority of last.fm users only have a few friends:

All this data was released on the Anonymous Prof's blog yesterday, where you can find even more images and details. Next up, he plans on collecting the listening history of users in order to examine music listening patterns as they relate to the network. Visualization junkies will want to stay tuned for that.

The non-profit organization called Sunlight Foundation, whose mission is to use the Internet to make information about the U.S. government more accessible,  just released a visualization of campaign contributions from 1990-2008, broken down by industry sectors and party lines. From this app, profiled on Programmable Web, you can see how the finance, insurance, and real estate industries spend more than others. The visualization is interactive - just push the play button after configuring the settings. It was built using Google Motion Chart and data from OpenSecrets.

2) Visualizing Earmarks

Earmarks are a hot topic in the current U.S. Presidential election. You can visit  the web site earmarkwatch.org to investigate those spending measures inserted by members of Congress into bills that direct taxpayer dollars to their pet projects. But an even easier way to track which states are the worst for using earmarks, this visualization over on ManyEyes is useful. Wow, look at Alaska!

3) Visualizing Election Polls

University of Utah computer scientists have written software they hope will eventually allow anyone to interactively and visually analyze election results, political opinion polls or other surveys. The software displays data in the form of "radial" charts that are doughnut-shaped and include features of traditional pie charts and bar graphs. The charts are interactive and animated, too. You can watch a video demonstration over here, but unfortunately, the poll-analysis software isn't quite ready for prime time. What a tease!

4) Electoral College Prediction Tracker

This interactive visualization widget provides an overview of the predicted outcome of the U.S. presidential election. The rows depict the results from different news agencies (The Washington Post, The New York Times, CNN, etc.) and the columns represent the different U.S. states. The states width is based on the number of electoral votes they have available. Political bloggers will really like this one, too - it's embeddable!

5) The 2008 Presidential Election In The Blogosphere

This next visualization, perspctv.com, is an informational dashboard that summarizes and graphs the Internet activity relating to the 2008 presidential elections. The charts compare the similarities as well as the differences between the mainstream media and user-generated content, such as that found on political blogs. Currently, the graphs include CNN polls, new mentions, blogosphere mentions, Twitter mentions, a U.S. electoral map, and Google Trends-based timelines. (via information aesthetics)

Since the Troasts started using the device six months ago, they've decreased their energy use by 15%. Now they're wondering if it will drop even more with the world watching their energy consumption online.

Developers, Want to Make Your Home Tweet?

Although not everyone will want to make their data as public as the Troasts have, we know there are probably a few "do-it-yourselfer" developers who have been waiting for a fun project like this to at least hold them over until Google gets their PowerMeter rolling.

Non-developers, try the Tweet-a-Watt system, instead.

Here's how to create your own real-time TED-tweeting home like the Troasts did:

Materials

• TED (The Energy Detective)
• ASUS WL 500G Deluxe wireless router
• OPEN Wrt operating System
• Google Visualization API

Methodology

1. You'll need a wireless router that has a USB Port and is capable of running an OPEN Wrt operating system. The Troasts went with the ASUS WL 500G Deluxe.
2. Replace the router's existing operating system with the OPEN Wrt operating system (Here's how.)
3. A customized script (customized software) has to then be put into the OPEN Wrt operating system. OK, so this is the hard part, Troast says the developers will post the script online at the EnergyCircle blog in a couple of days if there's interest. So if you're interested, please says so in the comments! Without this key piece of the pie, you're left writing your own script. (And if you do, share!)
4. Plug the TED into the wireless router using TED's USB port.
5. Write the ability to add an annotation to the data into the web site database so that you can annotate the data in the database.
6. Write the data display page using the Google visualization API, which enables you to take the data and annotations and make a chart like the one that you see on the EnergyCircle site. (Specifically, they used the Annotated Time Line feature in Google's visualization API).

Note - How the Data Streaming works:

• The script reads output from the TED every second and records it in a file located on the OPEN Wrt router.
• Each minute, the router posts the output into the Energy Circle Database.

About TED devices:

This year ReadWriteWeb gets a spot in the map, which we're all very proud of here in the RWW virtual office. You can see a full blow up of the image on Flickr and a zoomable version on zoomorama.com.

The flagship app is the Real-time Web Monitor, which shows the countries (or Canadian provinces or American states) that are experiencing the most traffic load. As I write this, California and the United Kingdom are together accounting for 15.5% of global data requests. The app also lets you view the ten worst performing cities (Hong Kong and Tokyo are really hurting right now), and the area experiencing the most outside attacks on the network. (Venezuela is not a safe place to be a server this afternoon.)

In addition to the three web performance visualizations, Akamai offers three performance indexes: News, Retail, and Music. They show traffic trends in different market sectors. For example, you can see the peak shopping times in different parts of the world. The retail graph also shows spikes around certain major shopping times of the year (there is a spike shortly before Mother's Day on the current five-month graph -- I'd expect a big spike later in the year right after Thanksgiving in the US: the official start to the Christmas shopping season here).

Above is a picture of the music index, but my favorite is the news usage visualization. The news index is neat because it has a list of the news events that caused the greatest spikes in traffic to news sites since Akamai started tracking in 2005. Interestingly, they're mostly all sports related. Taking the top spot? Ghana eliminates the US in the World Cup last June. Second and third is the first day coverage of the US college basketball playoffs for the past two years. The first non-sports news is the death of Anna Nicole Smith last February, which checks in at number four, and just 6 of the top 15 news-related traffic spikes are non-sports items.

Akamai also offers desktop widgets for its three net usage indexes. The widgets use Yahoo! Widget Engine for Windows and OS X's dashboard widgets for Mac.

Akamai's data visualization tools are attractive and addictive and a fun way to keep tabs on what's going on across the global net -- trafficwise, anyway. (Other net traffic visualizations include the Internet Health Report and the Internet Traffic Report.)