VizWorld.com

Robert Kosara (eagereyes) has written a fun little tool that posts random text from InfoVis papers at major conferences on twitter at the @InfoVis_Ebooks account.

Each tweet contains a reference to the paper the snippet is from. For InfoVis, VAST, and CHI, these are DOIs rather than links. Links get long and distracting, whereas DOIs are much easier to tune out in a tweet. If you want to see the paper, google the DOI string (keep the “doi:” part). You can also take everything but the “doi:” and append it to http://dx.doi.org/ to be redirected to the paper page. For other sources, I will probably have to use links.

It’s more humor than function, at first glance, but you never know when you might find one worthy of digging deeper. For example, the image above pointed me to an interesting paper on visualizing Ice Hockey Analytics, all because I wondered what “SnapShot” was.

via Meet @InfoVis_Ebooks, Your Source for Random InfoVis Paper Snippets | eagereyes.

Science eagereyes, infovis, twitter

Robert Kosara has a great piece on what he thinks will happen in the field of Information Visualization this year, and it’s a bit of a surprise.  It won’t be algorithms, it won’t be The Cloud, it won’t even be open data:  It will be Microsoft’s IE9.

But another development will actually be much more important: Internet Explorer 9. What has held back visualization so far is that no version of IE supports SVG or Canvas. Later this year, IE 9 will be released, which not only includes that support, but also has a hardware-supported implementation of Canvas. Practically all other browsers already support Canvas, but some are still behind on the performance part. That is likely to change very soon.

via The State of Information Visualization, 2011 | eagereyes.

Science infovis

I have followed Doug Short for a number of years now via some of his graphs posted on Calculated Risk. I especially love the fact that he has compared the crash of 1929, with the 1973 oil crisis, the 2000 tech crash, and the 2007 financial crash in a nice visualization.

Recently he has been taking a look at the growth in the Federal Debt, and comparing it with the tax structure in the United States. In the linear chart, we can see that debt is increasing, without bound it seems. He has also plotted a log scale version as well. There seems to be a disconnect between tax rates and the level of debt.

He then asks a very pertinent question. How accurate are the government’s forecasts about the future debt? If the government is accurate in its estimation of future debt, then we can trust these charts fairly well. However, if the government is usually off the mark, then we really cannot trust the forecasts for future debt. The resulting visualization is very illuminating:

These federal debt forecasts confirm we what already know — 2008 was a major economic turning point, a metaphoric fork in the road. However, the chart helps us quantify the magnitude of the new direction. The current 2015 forecast of a 19.68 Trillion debt is about 46% higher than the equivalent point (about 13.5 Trillion) on the road not taken.

via dshort.com: Federal Taxes and Debt.

Graphics infovis

I cannot stand a stacked bar chart for one simple reason. The stacked bar chart has only one common baseline. Thus you can only compare variables along that one common baseline. The rest of the variables do nothing except to clutter up the chart.

I have wanted to do a post against stacked bar charts and to suggest alternatives. Jon Peltier has beaten me to the punch, and has done a much better, and much more thorough job than I would have done. My favorite alternative was the panel column chart with series labels, while his favorite is the panel vendor bar chart. Really, I would call his favorite a bar chart matrix.

Take a look at his article to learn other alternative approaches.

This chart has another issue. The percentage items do not sum to 100%, so it is harder to understand what this tells us. 260% of Tableau users perform these eight tasks? What does that mean? It’s reminiscent of last December’s infamous Fox News pie chart.

I decided to examine some alternative charting approaches.

via Stacked Bar Chart Alternatives » Peltier Tech Blog.

Hat Tip to ChartPorn, which is where I found the reference to the article.

Graphics infovis

Hades: So you took care of him, huh? “Dead as a doornail.” Weren’t those your *exact* words?
Pain: This might be a different Hercules.
Panic: Yeah. I mean, Hercules is a very popular name nowadays.
Pain: Remember, like, a few years ago, every other boy was named Jason, and the girls were all named Brittany?

It turns out that Pain and Panic were right in the Disney movie Hercules. Stephen Von Worley has posted an interesting visualization of the most popular names for both boys and girls. It shows that in the late 1980′s Brittany was a very popular name, before fading. In the early 1980′s, Jason was also a popular name which has since faded.

Sorry Maitri, but your name does not show up on the list of girl’s names. But do not feel too bad, for neither does Randall or Paul show up on the list of boy’s names.

The rock of the girls names would have to be Elizabeth, and post-2000, Ava, Mia, Addison, and Chloe make a sturdy debut. Also, notice the recent retro comeback, wherein five long-lost favorites – Ella, Anna, Emma, Lillian, and Grace – have emerged from obscurity to reassert their vintage charms. Could Ida, Minnie, and Bertha be next? Time will tell!

via Name Change « Weather Sealed.

Graphics infovis

Did you know that there is less than 950 days (2.6 years) until regional registry IPv4 address exhaustion? There are 4,294,967,296 addresses (or about 28.8 adresses per square kilometer of land) for IPv4. Thus we need to moving towards IPv6 which gives us 2¹²⁸ addresses (or about 2.2 x 10²⁴ adresses per square meter of land). The TechRepublic has published an article and visualizaiton of the 2006 status of the blocks of addresses and their utilization. They even link to a XKCD map of the internet. From the article:

The IPv4 over-allocation comes up frequently in discussing the eventual exhaustion of available IPv4 addresses; in fact, Michael Kassner’s recent post explains how available addresses are almost exhausted and that the days are limited for IPv4. The current situation of the IPv4 address space can be visualized to show who has what addresses, what is full, and what is available.

via : Visualizing IPv4 addresses on the Internet

via : XKCD map of the Internet

Science infovis

Slate has taken the Local Area Unemployment Statistics from the U.S. Bureau of Labor Statistics to create an information visualization showing the jobs lost over the past two years. Blue circles indicate that jobs were gained. Red circles indicate that jobs were lost. The size of the circle indicates the magnitude of change. The visualization is interactive. It allows you to either play through the timeline, or you can click on a certain month to see the data. The problem with such visualizations is that entire states disappear from view under the data. I personally prefer the way Flowing Data used the counties in their visualization instead of using circles. Unfortunately, Flowing Data’s visualization is not interactive like Slate’s.

via Slate: When Did Your County’s Jobs Disappear?

Science economic, infographic, infovis, unemployment, us

USA Today has created an interactive visualization showing a jobs growth forecast from economic consulting firm Moody’s Economy.com. You can view the one year forecast change in jobs for the entire nation by state. When looking at the forecast by state, Texas looks to improve by 1.3% while Nevada continues to plummet by 1.9%. If the forecast is true, then Texas will improve from a November 2009 unemplyment rate of 8% to 6.7% in Novemerb 2010. Nevada will worsen from 12.3% to 14.2% over the same time period. Having 1 out of ever 7 people “officially” out of work (the U-3 number) is not good news for Nevada, and that does not count those who are discouraged and given up looking (the U-6 number).

You can view the one year forecast change in jobs for one of 384 metropolitan areas. Huntsville, AL should improve the most with a 3.2% decrease in unemployment. Vero Beach, FL will fare the worst with its unemplyment rate increasing by 3.6%.

On both the state and metro level, you can see how jobs will fare in one of fourteen different categories. The data is also given for each quarter out to the year 2013.

The problem with these information visualization graphics is that the headline does not match up with the data. The headline says “Jobs may rebound in 2010“. Yet when you take a look at the jobs forecast for the United States for all sectors, the data says that the job rate will worsen by 0.2% from November 2009 to November 2010. This means that the unemployment rate will worsen from 10% to 10.2% over the course of the year.

If you follow the unemployment rate, then another item to be aware of is the annual revision of jobs that occurs every January, and is scheduled to be announced on February 5th. The preliminary benchmark payroll revision is a minus 824,000 jobs. In other words, the BLS has been too optomistic on the unemployment rate over the past year.

This graphic shows actual job growth through third-quarter 2009 and Moody’s Economy.com’s forecasted job growth for fourth-quarter 2009 through third-quarter 2013. It covers every state, the District of Columbia and 384 metro areas, broken down by fourteen industry sectors.

But it could be worse. Spain has 19.3% unemployment.

via USA Today: Jobs may rebound in 2010

Science infovis

Robert Kosara, known to many as @EagerEyes, has written up his views on Information Visualization and access to data in 2009, and where he thinks 2010 is going.

In addition to the practical visualization uses, 2010 might be the year of visualization theory. While our field is certainly an applied one, we still need a much deeper understanding of how it works and how to build better tools. There is some existing work, but much of that is old (Bertin’s work was published in the 1960s, Mackinlay’s almost 25 years ago, Shneiderman’s 13 years ago, Chi’s taxonomy almost ten years ago). The field is progressing and we are developing new tools that do not always fit the old molds. We are also gaining a better understanding of how things work, and we are seeing interesting new concepts from other fields. So an update of our theoretical foundations is really overdue now, and this year will hopefully be when it happens.

It’s a great read, I won’t spoil it further.

via The State of Information Visualization | EagerEyes.org.

Science 2009, infovis, review

Do you ever wonder: “What the United State’s strategy for dealing with the insurgency in Afghanistan?” Do you ever wonder: “How do people come up with such bad visualizations?” Thanks to the PA Consulting Group, we get to see both at the same time.

Richard Engel, NBC News Chief Foreign Correspondent, obtained an unclassified briefing given to the Joint Chiefs of Staff. The briefing details how to conduct a successful counterinsurgency, or COIN.

… U.S. and NATO forces “must accomplish three tasks simultaneously”:

“Influence insurgent-minded individuals to adopt a neutral disposition.”

“Influence neutral-minded individuals to adopt a supportive disposition.”

“Retain supportive individuals.”

Accomplishing three tasks does not sound hard. But as always, the devil is in the details. It is one thing to come up with a vision. It is another thing to make that vision work. In steps the PA Consulting Group. They start with the three tasks, and over the next 21 pages in a PDF, show how to build a nearly incomprehensible visualization. The remaining 8 pages of the PDF discusses various details, such as reducing the territory that the counterinsurgency holds, winning the information war, and providing security for the population.

The slide is undoubtedly overwhelming. For some military commanders, the slide is genius, an attempt to show how all things in war – from media bias to ethnic/tribal rivalries – are interconnected and must be taken into consideration. It represents a new approach to war fighting, looking beyond simply killing enemy fighters. It underscores what those fighting wars have long known, that everything matters.

If you download the PDF and read it, please take a look at the presenter notes. While I think that many of the ideas they are trying to convey are good, how they present it is worthy of a Razzie. The PA Consulting Group is based in the U.K., they have a branch office based in Arlington, VA. My guess is that it is this branch office that created this monstrosity.

via MSNBC.com So what is the actual surge strategy?.

PDF via Afghanistan Dynamic Planning PDF

Graphics bad, infovis