June 13, 2011 Off

Clear Data on U.S. Tax Rates from the Center for American Progress

By in Neat Visualizations

Admittedly, I haven’t checked the “honest” half of this information, but Paul White’s article about US tax rates does a great job of clearly showing a number of different data sets.  Take this example, which describes income taxes paid by wealthy Americans:

chart of taxes paid by wealthy americans

Pretty much the only thing in the chart is information – a starting year and value for each line, and an ending year and value for each line.  The purpose here is to show the trend, not clutter the graph with meaningless data.  The artist takes the additional helpful step of eliminating a legend and fixing labels in close proximity to the lines (and even uses the same color for the label and the line!).

In the following chart, the artist makes use of both hue and intensity as preattentive attributes (pdf, page 5) to draw our attention to interesting data points (the smallest, the biggest, and similar countries in dark blue, the US in bold red):

Nice chart describing corporate tax rates

Notice how quickly you can understand the point of the chart – the United State’s position in a list of country corporate tax rates.

Overall a very nice, minimalist, and effective presentation of information.  Good way to end the weekend!


June 4, 2011 Off

Concept Design Studio: Playing with Sketchup

By in General Information

I’ve been playing with Google’s Sketchup product since before Google owned it (waaaay back in 2004).  It’s a fantastic product for creating quick 3D models, made all the more powerful by the library of objects created by Sketchup users.  I don’t use it often, but every once in a while it’s a fun way to pass an hour. Yesterday I did just that to imagine what my design studio of the future would look like.

About two weeks ago I was fortunate enough to get a tour of an industrial design studio in San Francisco.  The space was incredible — a building with hundreds of years of history housing some of the most fascinating product designs of the future.  The atmosphere was dripping with creativity, innovation, and potential.  It made the Google offices I work in, while fun and inspiring themselves, drab by comparison.

I had never been in a design studio before, and afterwards  I was asked how my preconceptions about design space differed from the reality. I struggled to put my thoughts into words, but the one component I was surprised at was the spatial separation of digital and physical design space.  To me, the two have become so intertwined that I believe it makes sense to combine them into a single workflow, so prototypes can be made on a bench next to massive displays showing renderings next to whiteboards with design sketches.  Instead of seeing digital efforts as a separate workflow, intertwine them physically with the method that has been used since industrial design has existed:  sketch, prototype, refine.

Unable to clearly articulate these thoughts, I thought it would be fun to spend an hour in SketchUp throwing together my future design studio (hah!).  Some of the highlights:

  1. The space is designed into pods consisting of computer space, workspace (big tables for messing with physical medium like foam board), creative space (huge whiteboards where sketches can be made or posted), and inspiration space (big LED televisions that can show renderings or cycle through photos that will inspire the designers).
  2. The designers’ computer space always faces the creative space, so they are continuously presented with the entirety of their efforts.  This allows for subconscious and conscious assimilation of the entirety of design information, and plays to the senses described in Daniel Pink’s “A Whole New Mind” – most importantly, design, story, and symphony.
  3. Windows.  Lots and lots of windows.  Sunlight makes us happy.  Happiness begets creativity.  Creativity inspires design.  Win.
  4. Some relaxing space to take a breather.  I am assuming the kitchen and meeting rooms are on a separate floor, so I’ve just tossed in a few inspiring design peaces (an Eames lounge chair, for example).

If I spent more time with it, I would do a few things differently.  First, probably lower the height of the whiteboards – I feel like the space has become a bit too closed off. Second, I might switch the external pods to be facing the windows – design should always be looking towards the world, right?  Finally, I should probably add some plants, and maybe a patio with a garden.  Nonetheless, it was really fun to see what could be done in an hour.  While it certainly won’t be winning any awards for beauty, it’s a really easy way to sketch some prototypes for your new office, house, etc.

So, what’s in your ideal studio space?




May 22, 2011 Off

Data Lite: Birth of Technology via Ngram

By in Neat Visualizations

Google’s Ngram Viewer let’s you see how often a word or phrase has appeared in Google’s scanned book library (full about).  Here’s a quick one with regards to technology over the last 170 years.  Kind of interesting how early on “telegraph” was mentioned, and what’s with “computer” showing up in 1900?

Google ngram about technology

May 19, 2011 Off

Dishonesty from the Wall Street Journal

By in Not So Great Information Design

The New Republic highlights the dangers of dishonest data.  Apparently the chart  on the left below, originally accompanying this article in the Wall Street Journal (the same publication that produced this book??), has been making the rounds of conservative blogs as support for arguments against increasing taxes on the wealthy – because, clearly, all the wealth resides in the middle class.  Right?

Mother Jones does us a solid by re-designing the chart (on the right below) in a slightly more (but not completely) honest fashion (I’m assuming this is due to their lack of source data).  The problem with the original, of course, is the arbitrary selection of bucket sizes, which range from $4,000 ($1K – $5K) to $4 million ($1M – $5M).  The presentation of data should always objectively represent the underlying information.  Simply using constant bucket sizes for the histogram (say, $10k buckets) would eliminate these issues.  Except in extreme cases, use constant bucket sizes for histograms.  Anything else skews data in unexpected ways.

Shame on the original designer and the Wall Street Journal for allowing this kind of crap on their site.

where the money is redesign by mother jones

May 18, 2011 Off

Use Tree Plots Sparingly

By in Experiments in Info Design, Not So Great Information Design

Big things stand out.  So, when Asymco decided to report on the relative market and profit share in the smartphone industry, they got big.  Really big:

The dark orange represents Apple, and the remainder of color-coding is meant to represent profitability – pink and orange are profitable, white is (assumedly) break-even, and blue is a company operating at a loss.  The design in the original post has a number of flaws that make it difficult to draw conclusions from the information:

  1. Chart choice.  This is the primary flaw of the design.  The use of a tree map-type style chart was incorrect for this data.  Though tree charts are capable of showing the relative size of different elements, they aren’t very effective.  Humans have a difficult time determining the difference in size between two areas, particularly when those areas that are not in close proximity.  For example, try to determine from the charts above who has a larger unit share, Motorola or HTC.  The rectangles are pretty much the same size, with Motorola possibly edging out HTC.  Tree maps also tend to take a lot of space to say very little.
  2. Color selection.  Two issues here.  First, the colors break conventional norms, with red representing profitability and blue representing a loss.  Typically, “in the red” refers to a loss, and black or another more positive color like green would be used to represent positive profits.  Second, there is no legend on the charts to describe which color means what.  White isn’t labeled anywhere, including the author’s explanatory text.
  3. Labels.  There is no reason not to label these large boxes with the full name of the manufacturer.  I could probably figure out what “SE” means, but I don’t want to (and neither does anyone else).  Also, what do “Diversified” and “Smart” mean?  I can pretty much figure it out, but I shouldn’t have to.
  4. Contrast and assumed proximity.  The contrast between the “Diversified” and “Smart” areas is low, and I have to assume that anything below the “SMART” label is a smartphone-only manufacturer.  Why should I make that assumption?  Tree charts don’t have to follow that convention, so the author is again taking liberties with his audience’s ability to interpret.

Here is a quick alternative, with estimated numbers since I don’t have access to the source data.  I’m not 100% sold on the side-by-side charts, but I thought I’d throw it out there and see if it stuck.  I also haven’t reduced the space of the chart as much as I’d like.  A better alternative might be a scatter plot, but my attempts yielded a fairly cluttered chart.  If I chose to reduce the information presented (like eliminating the distinction between Distributed and Smartphone-only manufacturers), a scatter plot might be very effective.

However, even in this simple redesign, notice how easy it is to draw conclusions both within a category (by units or by operating profit) and across categories.  I’ve eliminated the confusing color coding and used simple English to highlight companies without profits and to distinguish between the different types of manufacturers.  Color is only used to highlight the main point of the chart – that Apple makes a few phones but takes home most of the money.  What do you think?  How could this design be further improved?

Asymco Chart Redesign


May 12, 2011 Off

New Data Search Engine – Zanran

By in Tools


While in early beta, this is a pretty exciting place for a data junkie.  According to their About page:

Zanran helps you to find ‘semi-structured’ data on the web. This is the numerical data that people have presented as graphs and tables and charts. For example, the data could be a graph in a PDF report, or a table in an Excel spreadsheet, or a barchart shown as an image in an HTML page. This huge amount of information can be difficult to find using conventional search engines, which are focused primarily on finding text rather than graphs, tables and bar charts.

I tried it out on a few queries and it tends to be relatively PDF-heavy (with a few sprinklings of Wikipedia).  For example, check out this search for “NHL Statistics“.  The first result is a wikipedia page for Christian Ruutu (kind of a weird choice), and the remainder of the page has 6 PDFs, another wiki page (wikia), and two blog postings.  None the data is particularly generic (as you might expect for a search like “NHL Statistics”) and much of it is off-topic.  Compare this to Google’s search result, whose first four listings are the official NHL statistics site, ESPN’s NHL statistics, Yahoo’s NHL statistics, and hockeydb.com, a storehouse of hockey statistics.  While Zanran is meant to search structured data (which doesn’t exist on any of those links), I am guessing I would have far better luck starting my search on hockeydb.com than in any of the PDFs delivered by Zanran.

When I tried a more specific search, “US census data from 2000 – 2010“, Zanlan’s first result was a pdf of a presentation from Cumberland County, whereas Google’s first result led me to a link to download summary files from the 2010 census directly from census.gov.

I think Zanran has a great idea and I hope they succeed (despite my fear of this site propagating even more terrible infographics), but the early beta could certainly use some major improvements in relevance.  For now, I’ll stick with Google.

May 10, 2011 Off

Worst. Infographic. Ever.

By in Not So Great Information Design


Where to start?  Implied correlation?  Terrible pie charts?  Bad graphic design?  Misleading data?  Boring subject?


May 7, 2011 Off

Use color with intent; Eliminate moiré vibration

By in Experiments in Info Design, Not So Great Information Design

Just about a week ago, Apple released new iMacs with shinier, faster processors.  How much faster?  Primate Labs ran some benchmarks and put together the chart below to help us understand what we would get for our money.  Unfortunately they made a few mistakes in presenting that information too us.  Though pretty, their use of alternating colors serves no purpose – the new and old iMacs both end up being shown in both blue and green.  This makes it very difficult to understand the very point of the article – the improved speed of the new iMacs.

The chart also falls victim to the moiré effect – essentially, an optical illusion that makes the bars appear as though they are vibrating.  This distracts the viewer from the message – the data itself.

Minor annoyances include too many horizontal rules (viewers can easily identify positions in a group of  3 or 4) and some repetitive language (re-stating iMac at the beginning of each label).  Check below for my redesign, which does the following:

  1. Highlights new iMacs in a different color to make the chart’s main point obvious, and used that color repetition in the data labels as well
  2. Increases contrast between the title and chart, and increase the descriptiveness of the title
  3. Formats the data labels in a more readable number format (addition of commas)
  4. Shifted the data labels inside the bar to eliminate a mental connection for the viewer (and reduce the size of the chart)
  5. Eliminated some of the horizontal rules, and reduced the prominence of the remaining rules
  6. Eliminated repetitive language in data labels (‘iMac”)
  7. Eliminated the moiré effect from the data bars by using solid colors

Click for a larger version:


May 7, 2011 Off

Just a bit more clear

By in Experiments in Info Design

iSupply released its forecast for Mobile App store revenue through the year 2015, and did so in a very simplistic chart.  Could such a simple chart be improved?  I think so.  To begin, the chart does nothing to differentiate between actual app store revenue and forecasted revenue.  While this might be discernible from the range of years shown on the chart, it still takes time and effort to understand where the actual data stops and the forecast begins.  The chart also has a bit of additional ink that adds no value.  See below for the original and my quick redesign.

Redesigning an iSupply forecast chart

April 26, 2011 Off

National Debt: How did we get here?

By in Neat Visualizations

A nice, simple chart from NYTimes supporting this article.  Why do I like it?  It’s simple and it tells a story that supports the article.  It also cuts through the posturing to put the ownership for our debt on BOTH parties.  In short, it’s clear, and it’s honest (assuming the data source is represented accurately):

From NY Times - How the debt became what it is