Visual.ly analyzed the top 30 infographics posted on their site and determined that data visualization doesn’t matter:

Data visualization certainly matters when it comes to conveying information effectively, but when it comes to sharing, the answer is no: having data to represent is not a critical ingredient in infographics. More than half, or 53%, of the top 30 graphics do not contain data visualization. And by data visualization, we mean visual objects that are sized, colored, or positioned to represent numerical values.

I think what they actually mean is that data visualization is not the sole factor of a successful visualization. Since they are only analyzing the top 30 infographics, the minority 47% that had data visualization are still very successful. It would be a different story if the 53% of infographics without dataviz were the top successes and the 47% with dataviz were the bottom losers.

My hunch is that the successful infographics posted on Visual.ly are popular because, like other viral content, they strike a nerve, are of the moment, are humorous yet relevant, or have some other je ne sais quoi.

I’m a sucker for anything cute and bubbly, and the U.K. Energy Consumption Guide created by Epiphany is no exception. It combines a vertical scrolling site with a lot of data visualization about different types of fuel and how they’ve been used historically. Most of the charts are solid and the interaction adds an even higher level of clarity and understanding.

While I like this circle packing chart, I’m sure there will be doubters. It’s very similar to McCandless’ natural gas visualization that received a lot of flack. But generally speaking, anything that is engaging and welcoming garners a little extra time from the visitor to make sense of it.

I warned Nathan that I was going to drop a pie bomb on Flowing Data. Well, here it is; it’s labeled by its creator as a “feather chart.” I really hate to pick on people. I truly think Jon made a valiant attempt to use pie charts innovatively. However, this chart is not effective.

The chart uses 11 million ACT records (for international readers, that’s a standardized test in the US). It’s trying to show the relationship between ethnicity and test score and income and test score.

I created the y-axis as the ACT composite score, and then used self-reported income bands as the x-axis. Both are discrete, categorical values, even though ACT is numeric. ACT increases bottom-to-top, and income bands increase left-to-right. At the intersection of each variable is a pie chart, sized by the number of students in that group, and colored by ethnicity

The only problem is that the overlapping pie charts occlude one another. Unless one section of the pie chart dominates and allows the other sections to peek out over the top of the previous pie, then the chart is useless. For instance, in the first feather, there’s no way to know if the orange section is 40% or 60% for most of the chart.

This chart has really good intentions, but the data would be better served with a bean or violin plot. If you’re a subscriber, you can check out Nathan’s great tutorial from last week about visualizing distributions.

Gundega Strautmane, a Latvian textile artist and designer, visualizes social and physical networks in a show called Relational Ornaments. The networks are created using various sized pins to depict nodes and threads connecting them to show relationships. Bringing visualization into the tactile world lends it a weight not able to be achieved on a computer screen. It allows the viewer to pause, spend time with the information, feel it, sense it in a more holistic way. The placement of pins and threads is imprecise because they are placed by hand giving the work a very natural, organic feel rather than the rigidity of the exact calculations of programming.

[via The Network Thinkers]

This Wall Street Journal graphic shows who’s selling (or sold) a percentage of their Facebook stocks and who’s holding steady.

This graphic is the perfect example of why I’m a proponent of the pie chart. First, they stuck to two values per pie chart. That makes it easy to read. Next, they used the size of the pie to denote the number of shares. Finally, they used small multiples to easily compare both the shares owned by each entity as well as change in percentage of shares being sold.

I’m sure bar charts would be fine too, but WSJ really used all aspects of the pie chart very effectively.

[via Barry Ritholtz]

Washington Post’s Ezra Klein busts on the filibuster. Gone are the days of Mr. Smith when invoking the filibuster was seen to serve a greater purpose. The filibuster has its roots in Ancient Rome, and apparently even then it had its critics.

This chart is a great example of providing a lot of information in a concise area. All of these data points are relevant to the topic and helps us inform our opinion about the matter.

[via @hfairfield]

The New York Times does it again with this succinct look at tech IPOs. It begins with looking at everything through the lens of when Google’s IPO in 2004, which, at the time, was considered huge. The next screen adds Facebook to the mix which dwarfs everything prior. It continues on to show the first day of trading pop and where things landed long term (3 years post-IPO).

It’s a very interesting view of IPOs and could actually be a good financial analysis tool with a few more features.

Help is a drug company that offers you less. Less active ingredients, less waste, less confusion, less greed. Its tongue-in-cheek website has a map of its latest sales data called “What’s wrong U.S.?” A bar chart for each state shows how many people are buying products for particular maladies.

So why are the inner northwest states having problems sleeping? My guess they’re up late worrying about gay marriage.

Garbage in, garbage out the old adage goes. Nigel Hawkes, Director of Straight Statistics, describes a sort of statistical whistleblowing letter to the British Medical Journal.

A team from Imperial College found that in 2009-10, nearly 20,000 adults were coded as having attended paediatric outpatient services, and 3,000 patients under 19 were apparently treated in geriatric clinics. Even more striking, between 15,000 and 20,000 men have been admitted to obstetric wards each year since 2003, and almost 10,000 to gynaecology wards.

It’s hard to put your faith in analysis, visualization, policy, and anything else that comes out of data with reports like these. With human error being a known issue, we have to find better ways of inputting and double-checking data. Unfortunate mistakes at the outset only lead to bigger problems down the line.