Archive for the ‘dataviz’ category

Dataviz of the Week: Map It Yourself

3, October, 2008

FortiusOne, a Washington-area digital mapping company, has released something called Maker!. It’s a roll-your-own data visualization tool that allows you to mash up maps with a database and produce something that makes Google Maps look like they were produced by a computer running Windows386.

Below is a Flash-based map that takes a database of funding that’s gone to tech companies based on size and location. You can instantly see where the money is going.

The data is entirely play-friendly–zoom into geographic areas, choose big or small investments, etc.

Here I’ve zeroed in on the D.C. area, where I ply my trade. A popover shows me that one of the big recipients of capital in my home town is Clearspring, the nation’s lead widgeteer. [Huh. I wonder if they are hiring extremely good natured, value-priced, virtually hairless web consultants.] But it also reveals dozens of digital businesses I’d never heard of–GeniusRocket, Acumen, Price Comparison Guru, Brainware–within a 20-minute commute of my home.

The site is easy to build with [though beyond the skills of rank amateurs]. And it’s already got a gallery full of some fascinating stuff.

Let’s say, for instance, that you’re a political operative trying to target women aged 18 to 30 to vote in November. Bam, you’ve got a map that breaks down this population down to county and city level, all across the U.S.

Of course, such a map could have multiple uses. Let’s say you’re an unmarried man aged 18 to 30 wondering where the numbers are most likely to work in your favor, potential-mate-wise.

Gentlemen, start your engines.

Dataviz of the Week: Failed States [Other than Ours]

10, September, 2008

As we brace for the hysterical doom-and-bloom rhetoric of the general election, what better time than now to explore cases of real national failure and success?

The image above is a datavisualization of The Failed States Index, a report co-published by the Fund for Peace and Foreign Policy magazine. It evaluates 177 countries in terms of how close they are to, well, failure. [More on this below.]

As I have confessed repeatedly here, I’m a big fan of “dataviz,” as it’s known in the trade. Data visualizations demonstrate the power of images to illuminate information in ways that words alone cannot. I think journalists, educators and all professional communicators ignore dataviz at their peril.

Anyhow, the Failed States map is pretty simple as these graphic explainers go. The work of a map geek who goes by the handle “Ender,” the dataviz essentially turns each country’s failure index number into a color, allowing you to eyeball the places on the world map where countries are teetering on the edge of national catastrophe.

The visuals force fascinating questions to mind:

  • Is it significant that so many states near failure are located near the equator?
  • Why do nations seem to be stabler the closer they are to the North and South poles–with the glaring exception of Russia?
  • Why makes Ghana so much more stable than Guatemala?
  • What measures of national stability rank Portugal above the U.S.?
  • Why are China and Russia closer to failure than Cuba?
  • What happy sauce do they drink in Chile that makes that nation as stable as our own?

Which brings us back to the underlying data.

The Failed State Index is a calculation based on information about each country regarding 12 criteria, a research-and-analysis process that’s been vetted and validated by multiple layers of academics and globalist wonks.

Measures of national stability accounted for include Legacy of Vengeance-Seeking Group Grievance or Group Paranoia,Uneven Economic Development along Group Lines,” Suspension or Arbitrary Application of the Rule of Law,” “Widespread Violation of Human Rights,” “Progressive Deterioration of Public Services,” “Rise of Factionalized Elites,” and “Sharp and/or Severe Economic Decline.”

Maybe I’ve been following the presidential race too much, but this sound a lot like the talking points of the guests on both MSNBC and Fox News.

Obviously, people in this relatively stable nation-state of ours are very polarized over the forthcoming presidential election. I’m already hearing people recite the common refrain, “If [the other guy] wins, I’m moving to Canada.”

But why choose our neighbor to the north, which is hardly more stable than Austria, for god’s sake?

Using the handy Failed States datavisualization, it’s easy to see that if you’re looking for a rock-solid haven free of political instablity to sit out an unbearable presidential administration. . . Norway is the place to go.


Bookmark and Share

RealClearPolitics: Winning the Digital Journalism Race

29, August, 2008

Not long ago I castigated Congressional Quarterly for presenting high-quality reportage on political polling via a blog. They’re missing a great journalistic opportunity–to present daily analysis of the latest state-by-state Obama vs. McCain polls in a way that takes full advantage of the interactive visual medium that is the new platform for journalism.

It’s a classic case of old media not understanding what to do with their great stuff. Failing to “unlock the value” of their work, as they say in the corner offices.

Anyway, I’ve since discovered that such a map–a dataviz, or datavisualization, in web argot–exists. Unsurprisingly, it’s the work of a new media firm unburdened by an analog heritage.

The map is produced by RealClearPolitics, an online-only political analysis operation.

The map is a thing of digital beauty, a tool that lets you dig into good polling data smartly analyzed and interact with it by imagining various scenarios.

What if current polling holds through November? [Results shown above, pre convention "bounce."]

What if Obama wins Virginia and New Mexico and the rest of the ’04 results are unchanged? [Obama wins by a hair.]

What if McCain sweeps the rust belt of Pennsylvania, Ohio and Michigan? [McCain by a mile, even if he loses Florida, etc.]

You can base all of these scenarios on the latest polling data so you can see how realistic your own speculations are.

It’s great work, a simple dataviz that presents best-of-class information in a fully interactive way that delivers a very high level of public service. It’s “civic engagement” on a screen.

If old media doesn’t start winning this kind or race soon, there will be no doubt who will carry the contest for the media future.

No matter who the President is.


Bookmark and Share

Worst DataViz Ever: CQ’s Poll Tracker

13, August, 2008

I often write about great datavisualizations–applications that use interactive graphics to illuminate a database in inventive ways. A great dataviz explains stuff in a way words alone cannot.

Today I’d like to pay tribute to one of the worst data presentations of the political season: Congressional Quarterly’s Poll Tracker.

Let me say first that it’s a great idea to take the most recent state-by-state presidential polling data from the most credible sources and update it daily. Put some experienced reporters on it so they’re not fooled by bogus numbers. This will produce an electoral map showing the latest polls in all 50 states. What more could an obsessive horse-race watcher ask for?

Unless you decide to just report the data in a blog, without connecting it to a map, and just leaving it in the order that the data comes in. Here’s what you get:

I thought this presentation looked eerily familiar. Then I recalled the two-year mobile broadband service contract I signed over the weekend. You know the way they print out those contracts on long receipt tape? And they have to fold it over four times just to get it in the bag? That’s what the CQ “dataviz” reminded me of.

This is a classic case of journalists not understanding that how you present data is just as important as the underlying data itself. Stick that daily-updated state-by-state polling data on a map, float the data on flash pop-ups and you have a powerful application, a real reader service and eyeball draw. Leave it in a blog and all that reporting. . .turns invisible.

To be fair, CQ does have projection data on a map for House, Senate and Governors races. It doesn’t appear to take the most recent polling data into account, but it toggles neatly between current landscape and projected election outcomes.

Oh, wait, look! There is a “President” map that presents the latest polling data! My mistake!

Oh, never mind. . .that’s the results from the 2004 election.

Their mistake.


Bookmark and Share

Dataviz of the Week: bubbl.us

7, August, 2008

I’ve often yammered about how the rapid development of datavisualization–things as simple as timelines, as nutty-cool as you see in Digg Labs or as brain-stretching as in the gallery at Visual Complexity–will be culturally transforming.

By allowing people to see relationships dynamically, over time and in multiple dimensions, dataviz tools surface new understandings and ideas invisible via words and images alone. It will help us see obvious stuff that’s been hiding right under our cerebral cortexes for centuries. The rise of dataviz technology will unloose the vast intellectual capacity of people who think visually and spatially but maybe aren’t so good with words or numbers.

But there I go, yammering again. I came across a new dataviz tool that lets you brainstorm visually. It’s fun, it’s functional, it permits easy collaboration. It may change the way I handle projects and problems.

bubbl.us is a tool that lets you plank out ideas and chunks of information in “bubbles” and show how they relate to each other. It’s essentially a dynamic, flexible whiteboard.

It’s a freeware version of something called “brain-mapping” software, the expensive forms of which are used by businesses, universities, consulting groups and other serious thinky outfits with big budgets and high-stakes projects.

By allowing you to “see” ideas and how they relate to each other, it supercharges the brainstorming process.

Here’s the result of about 15 minutes of cogitation about the launch of a new product I’ve been working on. [I've left it tiny since it involves a real company.]

This is nothing exotic to look at. But I’ve done hundreds, maybe thousands of these brainstorming things on paper or whiteboards, and this is better. bubbl.us captures ideas quickly, lets you move stuff around endlessly and accommodates those “oh, yeah, I almost forgot!” and “hey, what if we just. . .” brainstorming moments.

Stick something in the wrong place and then move it. Kill a bubble and a cool little puff of smoke pops up.

When you suddenly see connections, you can draw lines or simply re-position a group of bubbles.

Bubbl.us is a work-in-progress, still funded with PayPal donations with a new beta expected soon.

There are certainly competitors. Exploratree invites you to use thinking templates, but they strike me as PowerPointlike braintraps.  Mindmeister makes me work too hard and is very word-heavy, but has lots of features (and a 2.0 version imminent).

Meantime, I like bubbl.us. I’m hoping it liberates my own vast, undiscovered intelligence and unlooses it upon the world. But I’m guessing it’ll maybe help me brainstorm more efficiently. That’ll do.

Crowdsourcing Crime: UCrime.com

5, August, 2008

Geographic visualizations of crime data are already old hat. At least since 2005, when peerless journogeek Adrian Holovaty created chicagocrime.org, people have been mashing up public crime data with various maps to illustrate where, in a manner of speaking, the bodies are buried. [Chicagocrime.org has since been swept into Holovaty's latest adventure, Everyblock.com.]

UCrime.com, a Baltimore startup launched last month, takes crime mashups to college, providing visual reports on incidents on over 100 college campuses. The picture is not always pretty. Here is a snapshot of the last six months of mischief that’s taken place at the University of Maryland at College Park, the school my son will be attending in the fall:

Looks like those crazy Terps have a blast on campus, doesn’t it?

Meanwhile, take a look at Brigham Young. Crime? Not so much.

The icons are kind of humorous (unless of course you’re the victim of one of the incidents). A spray can shows malicious destruction of property, a moneybag is theft, a fist a simple assault. Handcuffs show a successful collar. Users can choose to view all crimes or just, say, burglaries.

While the site is just launched, it promises to introduce a couple of social media features. It appears students can join a sort of digital neighborhood watch and report crimes. Users can “comment” on specific incidents or collaborate like junior crimesolvers.

Crowdsourced crime reports, “reviews” of certain incidents, collective responses to crime. . .Call me a worrywart, but if I were running this site I’d want to have a skilled moderator–and an even more skilled lawyer on retainer.

It’s worth noting that there’s nothing new to the information here. Campus newspapers always run crime reports. Local cop agencies make this material public. UCrime simply collects the information over time, tags it by type and connects the crimes with geography.

But it’s a good illustration of the power of even a very simple data visualization. The medium transforms a public datastream into a compelling story about a community and what goes on there.

Of course, that story is misleading. Three top-of-the-head reasons:

  • A compact campus with a given level of crime looks more crime-dense than a spread-out one.
  • The visualization does not take into account the size of a student body–“there’s no denominator,” as they say in applied stats class.
  • A quick glance makes it hard to distinguish a campus where there are dozens of open-container violations from one with a lot of gunpoint robberies.

In the end, perhaps the most valuable service is the one that lets students get alerts–via mobile phone, if they like–of crimes occurring within a specified chunk of geography. It’s good to know two kids just had their laptops taken form a certain dorm, for instance.

Of course, there’s nothing keeping a parent from signing up for this service too.

Yikes. What dorm is my kid staying in this fall again?

The 2.D’oh! Roundup: Oldpapers, Winning Money and McCain in Plain View

25, July, 2008

The Print ‘n’ Read Feature

This week’s Print ‘n’ Read feature–my recommendation for an online article so worthy that you might actually want to print it out and read it offline–is rich with irony. It’s a 74-page PDF about the future of journalism–as seen by the people who are running newspapers.

It’s tempting to dismiss this report, from the Project for Excellence in Journalism, with a consider-the-source wave. But the report ["The Changing Newsroom: What is Being Gained and What is Being Lost in America’s Daily Newspapers?"] is based on a thoroughgoing study based on face-to-face interviews and legit-survey-style questionnaires sent to newspaper leaders.

It’s at least an intellectually honest attempt by journalists to assess what’s happening to them. As a result, it’s less idiotically defiant and self-serving than many similar efforts. The people surveyed seemed downright chastened.

I’ll spare you the details, but it’s full of stories of optimism and ambition and furious attempts at innovation, all against the backdrop of a breathtaking descent into financial ruin.

My favorite oddball gem, so sweet and earnest and foolish you just want to pinch the cheek of whoever thought of it and send ‘em to bed: Some unidentified newspaper tried to sell copies of its important investigative report on Amazon.com.

How to Thrive in a Down Economy, Part LCVII

I love playing with ComScore’s news releases. Everybody pays attention to the top of the list to see how the Big Dogs are doing. I like scouring for other details.

Like this latest, from a list of Top 10 gainers over the last month. With a 30-day rise in traffic of 409 percent, the entry at the top is. . .GSN.com, home of the Game Show Network.

Why the spike of such an inane property? Always hard to tell. But it’s worth noting that the economy’s lousy, and GSN gives out cash prizes. And its latest sweepstakes? You can win a $500 gas card.

Dataviz of the Week: Partisanship in Plain View

On the impossibly-cool datavisualization site Visual Complexity I found this gem, Voting Patterns Among U.S. Senators, which depicts voting relationships among U.S. Senators in 2007. It was created by the Human-Computer Interaction Lab at the University of Maryland and published this spring.

The graphic demonstrates that Democrats tended to vote as a herd in 2007. The GOP? Not so much.

And why is the senior Senator from Arizona hanging out in the middle, unaffiliated, with Sen. Sam Brownback? He was campaigning for President and didn’t vote much.


Follow

Get every new post delivered to your Inbox.