http://blogcritics.org/ A sinister cabal of superior bloggers on music, books, film, popular culture, politics, and technology - updated continuously. en Copyright 2005-2007 by the authors Mon, 10 Oct 2005 09:51:59 EDT http://backend.userland.com/rss Blogcritics.org custom software http://blogcritics.org/ Phillip WinnSunday, August 26, 2007, marks the switch of all Blogcritics.org article feeds from full-content to short-content. This is the result of several converging factors, and is unfortunately a permanent decision (as permanent as any decision can be on the web, that is). We are aware of all of the reasons that this is a Bad Idea, and we are aware that some of you will be quite upset about having to click on something to read the free content, and we're sorry. Unfortunately, despite great effort, full-content feeds are not currently economically viable. Two other factors are involved: full-content feeds have resulted in an unprecedented level of content theft, with BC content appearing on many websites, usually spam sites, without attribution or permission. This duplicate content causes a cascading set of problems, not the least of which is that search engines generally aren't favorable to duplicate content, and don't always guess correctly. Finally, our RSS advertising partner is strongly in favor of short-content feeds. We hope that you'll continue to subscribe to BC via RSS, and when an article grabs your eye, it's only a click away, still free on the BC website. Thank you for your understanding. Administration0@blogcritics.org Sun, 26 Aug 2007 12:00:00 EDT http://blogcritics.org/archives/2005/10/10/095159.php KaiserIncome distribution is often presented in "Lorenz curves". The following is an attempt to defy conventional wisdom: A few elements of this chart are confusing. The same set of shading is used to classify two different variables. The annotation of "Top" and "Bottom" appears arbitrary. The rightmost column, representing top 0.1% of taxpayers, has the same width as the leftmost column, representing the bottom 20%. Why not stick with a Lorenz curve? This presentation is versatile. The diagonal is a line of "equality": for example, the (20%,20%) point on this line indicates that the top 20% of the population (ranked by decreasing income) took exactly 20% of the income growth in 2003.A lot of information can be read off this chart: the top 0.1% earners took about 25% of all income increase; the top 1% took 40%; the top 20% took almost 80%. In general, a curve that bends away from the diagonal (like this one) depicts severe inequality.However, in many practical situations as is here, comparison with the diagonal is meaningless. The "ideal" society would probably not be one in which annual income growth is equally distributed among all taxpayers. (I'll leave the rationalization to economists and sociologists.) More helpful is a graph that shows relative changes in inequality. If I add a second curve (orange) showing the distribution of income (as opposed to income growth), then we see a trend of increasing inequality! The large bend from the diagonal indicates that income distribution is far from "equal"; what's more, the distribution of incremental income is even more skewed.For more analysis of graphs in the media: Junk ChartsReference: "At the Very Top, a Surge in Income in '03", New York Times, Oct 5 2005.Ed/Pub:LisaM Culture37656@blogcritics.org Mon, 10 Oct 2005 09:51:59 EDT http://blogcritics.org/archives/2005/08/31/055227.php KaiserDavid Leonhardt re-opened the debate about whether high-spending baseball teams (like the Yankees) are winners or losers. According to his application of an idea from Doug Pappas, George surely fools his investors! Accompanying his article was a table of numbers, of which I clipped the top third: As tables go, this one is fundamentally sound, teams sorted by "cost per victory" which was the point David wanted to make.If some readers find this table hard to swallow, they probably have wandered off, trying to make sense of the payroll and winning percentage columns; or perhaps they got dizzy trying to get their heads around 1,133,807 versus 1,225,575. Precision is a great scientific virtue but rarely makes a good graphic guideline.This set of data, essentially a bi-variate series, gives me yet another opportunity to discuss the versatile scatter plot. Here is the basic design, with winning % on the y-axis and payroll on the x-axis. Contrary to the article's conclusion, there appears to be a general association between payroll and winningness. The dotted lines are median payroll (US$ 63 million) and median winning % (0.500) respectively so that half the teams fall on either side of each line. I have removed the Yankees since its spending far outstripped every other team (will return to them later).We can take this design a step further by standardizing both variables: in the new graph, the scales are in units of standard deviations (s.d.) so that 0 is the mean payroll and +1 is payroll that is one s.d. above the mean and so on. Observe that the Yankees payroll of US$ 206 million is four s.d. above the mean payroll.Notice the rectangle above. These are what I call "middle market teams", their payroll within 1 s.d. of the mean, ranging from US$ 39 to 107 million. Plotting them separately from the Big/Small Spenders gives us a much richer picture of what is occurrring in baseball today. On the left, the 25 middle market teams are almost equally distributed among the four quadrants (about 6-7 teams in each), showing possibly payroll having nothing to do with winning. However, extravagant teams (Yankees, Red Sox) always are winners and miserly teams (Pittsburgh, Kansas City, Tampa Bay) always are losers, the inevitability starkly revealed on the right. (Admittedly, these sample sizes are small.) Scatter plots reveal many more insights than tables of numbers. Any table must be sorted in one given dimension, and such ordering causes difficulty in understanding other variables listed in the same table. In a scatter plot, both variables are accorded equal status and the reader decides where to place her attention.For more analysis of this data, visit Junk ChartsReference: "Passing on Blue-Chip Players can Pay Off", New York Times, Aug 28, 2005. Sports35130@blogcritics.org Wed, 31 Aug 2005 05:52:27 EDT http://blogcritics.org/archives/2005/08/20/210442.php KaiserWhile reading Howard Wainer's Visual Revelations, I stumbled on the science of armouring our troops, through Howard's explanation of the following gem of a chart.Imagine you're an engineer working for the military. You have the ingenious idea to inspect planes that returned home and plot the pattern of bullet holes. The dark regions had high density of bullet holes. Your task is to recommend where to put extra armour on the new planes. What would you recommend? (Note: the answer appears after the graphic!)Howard credited Abraham Wald for his counter-intuitive insight. We should put extra armour in the white regions, not the dark regions. The inference is that the planes that got shot in the dark regions managed to return to the base while others got hit presumably in the white regions and never returned.What has this to do with sampling? If we forgot about the planes that never came back, we may jump to the conclusion that we should reinforce the dark regions. The sample we didn't see is as important as the sample we observed. To wit:Statisticians call this "survivorship bias". We only oberve survivors but we must not forget about the non-survivors!For more commentary on graphs and charts: Junk ChartsEd/Pub:LM Sci/Tech34523@blogcritics.org Sat, 20 Aug 2005 21:04:42 EDT http://blogcritics.org/archives/2005/08/16/101706.php KaiserWhile on holiday, I picked up this interesting chart on the Argentine paper Clarin, showing the results of a (daily) on-line poll. This presentation of percentages as a dot matrix of 100 points is the same concept that I described before as a "decile plot". Not sure why but at clarin.com, a different graphic (the bar chart) was used to present the same data. The bar chart is perhaps most visually appealing but the dot matrix allows readers to read off actual percentages by counting off the dots.For comparison, look at the pie and donut charts that many publications would no doubt choke us with if they get their hands on this data.I have already voiced my distaste for pies and donuts, here and here. OK, so they tell us the percentages add up to 100 and the bar chart doesn't. But how important is that factoid? So I still say: never use pies or donuts.The dot matrix/decile plot has some potential but I'm not sure if it is better than the bar chart. Here is one possible rendition:I would rather not tip the square. (I also dislike the color scheme but have not altered it.)Since I don't care to tell readers they add to 100%, I stacked the groups one over the other so that it is much easier to eyeball the exact percentages. Because of this, I can omit printing the percentages. Of course, now the onus is on the graphic designer to make sure there are 100 dots, no more no lessTo reconcile form and function, I left off the decimal point. When you plot 100 dots, you have made the decision that each 1% is important so why would you then print 47.6% rather than 48%?The irony is that for one data series, just printing the table is as good as anything. Culture34221@blogcritics.org Tue, 16 Aug 2005 10:17:06 EDT http://blogcritics.org/archives/2005/08/04/091934.php KaiserThis graphic, from NYT, was produced for admiration. Lets stare at it for a minute.Undoubtedly an uncommonly beautiful map, I rate it below the previously praised Wi-fi Nation map. What renders it so visually alluring is its high resolution, what Tufte calls a high data-ink ratio. Indeed, there is almost no wasted ink, every pixel carrying some data. If we reduce the resolution, the map will for sure look a lot less impressive. So this is one of the few instances where I'd allow such high data density.Why is this short of perfect?By itself, the map does not convey much insight. If we ignore the text and the legend, pretty much the only message we can read from this map is a tale of two regions: the eastern half is heavily human-influenced, and most of the less-influenced land is in the west. We will be misled into thinking that the red and the green each takes up roughly half the area. We can see highways and cities but that is not telling us much. If carefully examined, we can make out the Central Idaho Wilderness.In other words, all the important insights are conveyed via add-ons such as the call-outs to the 4 most pristine areas of the country. Also the information-laden legend, magnified here:Through this legend, we learn the amount of wilderness. Observe that red only constitutes less than 20% of the country yet we thought it covered half of the map.Make no mistake: this is an extraordinary legend that is a graphic in its own right, a bar chart showing a distribution.What separates a good graphic from a great graphic is the union of form and function. Here our brain can pick up region-level cues (eastern half vs western half) but the map provides us highly detailed data (municipality-level?), which contribute to the beauty but not our understanding of the data.Reference: "Where the Human Footprint is Lightest", New York Times, July 31 2005. Culture33635@blogcritics.org Thu, 4 Aug 2005 09:19:34 EDT http://blogcritics.org/archives/2005/08/03/001808.php KaiserHere is one terrific map, courtesy of Ray Vella at Business Week. Great job!This map works on two levels: The red and green dots provide strong visual cues to support the conclusion that Wi-Fi networks are being widely deployed across American cities, except in the mid-west The three shades of brown show the number of networks installed or planned in each state. Inclusion of such state-level information justifies the printing of state boundaries. Without plotting state-level information, state boundaries become chartjunk, as in the heat wave maps I've previously discussed. What's more, we can assimilate the city and state levels. For example, focusing on Texas, we see from the dark brown shade that it is a state with many networks, and then from the dots, we can see further where those networks are.A few minor improvements could be made: Tell us the upper bound of the legend, and name the legend. By changing 10+ to 10-X, the designer not only provides us another piece of data but also harmonizes the presentation with the other two categories. Besides, the legend needs a title Be more friendly to the color-blind: the red-green contrast should be avoided as much as possible. If a graphic designer is reading the blog, please tell us where we can find studies of color contrasts Use a softer national boundary: the solid black line sticks out against the soft background and it is the least important bit on the map One would expect the choice of three shades of brown and the intervals used for each shade to be keyed to the frequency histrogram of the number of networks (shown right). The current division divides the 50 states into groups of 8, 16 and 26. Are there better divisions? Finally, most readers will find the number of networks to be a dissatisfactory metric because more populous states will likely install more networks. A density measure such as networks per person or per household or per unit area would have been more telling.Reference: "Wi-Fi Nation", Business Week, Aug 1 2005, p.12. Culture33567@blogcritics.org Wed, 3 Aug 2005 00:18:08 EDT http://blogcritics.org/archives/2005/08/01/053420.php KaiserSeveral readers requested that I explain and extend my last posting on the Global 10 companies chart printed in The Economist (see right). The primary distortion in this graphic is the unequal treatment of two data series, revenues and profits. The decision to treat revenues differently from profits is arbitrary and distorting. The exact values of revenues are missing, being coarsely represented by horizontal bars against a grid while the precise values of profits are printed to 1st decimal place but with no visual aid for comparison. The alternative graph I suggested provides equal footing to both series of data: Here, data values and horizontal lines are plotted for both revenues and profits. Sizing up the profit lines, we can compare profits between these 10 companies: this is hard to do with a list of numbers. As previously stated, the chart on the right using lines rather than thick bars is preferred: perception studies show that we are better able to compare lines than bars because we do better at comparing line widths than areas. Notice that the height of horizontal bars (similarly, width of vertical bars in a column chart) does not encode any data.As a further enhancement, I'd make the profits side of the chart roughly equal in size to the revenues side. This feature may mislead some to think that profits and revenues are roughly equal in magnitude but it helps those who want to compare profits between the companies. A much better chart, favored by statisticians, is the following scatter plot, which brings out insights from the data. The 10 companies fall into three groups: Giants (Exxon Mobil, Royal Dutch/Shell, BP, Walmart), Big and Mean (GE, Total, Toyota) and Big and Fat (Ford, DiamlerChrysler, GM).Giants have revenues over $270 billion, among them the oil companies are more profitable than Walmart, even though it has the highest revenues.Big and Fat consists of auto companies with big revenues but low profit margin.The ranking and relative size of revenues can be read off the horizontal axis. Similarly for profits on the vertical axis. So both bivariate and univariate distributions are available on one chart.This chart is not without problems. It is difficult to place the data labels without obstructing readability. I placed the labels where they don't interfere with any reader tracing dots to the axes, effectively in the northeast quadrant of each dot. I omitted the country labels so as not to clutter the space. I used color to further separate the dots from the labels. I also sacrificed printing every data value for readability, assuming that the reader's interest in magnitude is not absolute. pub:NB Culture33447@blogcritics.org Mon, 1 Aug 2005 05:34:20 EDT http://blogcritics.org/archives/2005/07/31/093552.php KaiserThe chart on the right, from the Economist, displays revenues and profits of the 10 largest companies in the world in 2004 (as ranked by revenues). The graphic is yet another container of data: it elevates revenues, treating profits as a side thought. This table is clearer: The next two charts attempt to put revenues and profits on an equal footing. I include the left chart to illustrate the folly of typical bar charts: the height of the bars serves no purpose except to distort our judgement of the lengths. The right chart is preferred. But there is another graphic, favored by statisticians, that truly brings out insights from the data. Go to Junk Charts for more. Reference: "The World's Biggest Companies", Economist, July 28 2005. Sci/Tech33405@blogcritics.org Sun, 31 Jul 2005 09:35:52 EDT http://blogcritics.org/archives/2005/07/29/080001.php KaiserI have never come across a situation that calls for a pie chart. The human mind thinks linearly: we can compare lengths of line segments but when it comes to angles most of us can't judge them well.The donut chart is a pie chart with a hole punched in the middle. Alas, the missing middle contains the angles that help us size up the slices. The donut chart is a useless chart made worse. Never ever use a donut chart. Each publication gravitates to certain "pet" charts: The Economist happens to like donut charts. Hopefully their editors will read this and stop using them. Here is a recent example: We might as well point out three additional crimes: firstly, having one donut as a mirror image of the other denies us any chance of comparing like-colored slices properly; secondly, the lines linking labels to slices positively make us dizzy; finally, the least important detail, i.e. the total population size, stares us in the eye.Reference: "The Americano Dream", The Economist, July 14, 2005For reviews of graphics in the media, visit Junk Charts. Published: NB Sci/Tech33299@blogcritics.org Fri, 29 Jul 2005 08:00:01 EDT