Data Visualization & Graphs

For my annual April Fool’s Day post, we revisit some of the past year’s most useless and inane numbers presentations. This year we celebrate an infographic about marijuana use in the U.S. in the recent past. This graph is this year’s winner because it is not only shockingly incompetent, but appeared on the website of a news outlet – CBS News – that really should know better.

Hans Rosling, age 68, died earlier this month. He contributed more than almost anyone to helping people truly understand the world as it is. In an era when we have gone overboard for the glitziest data visualization techniques – frequently at the expense of comprehensibility or of any meaning whatsoever – Rosling’s presentations are informative but still imaginative and entertaining. His work demonstrates the potential that data visualization can have, but all too often doesn’t – we’ll see one of his best examples here.

My last blog looked at how U.S. federal government outlays were distributed in 2015 – one of the few situations I’ve found where a pie chart actually works! But taking a deeper look at the issue, and especially considering federal spending trends over time, we’ll see that pie charts are severely limited. We need a better presentation approach, and in the process we’ll come face-to-face with the limitations of data visualization.

You know how much I hate pie charts – in my book, "Painting with Numbers," using a pie chart is Deadly Sin #10. But occasionally – very occasionally – a pie chart tells an important story. Current federal government spending is such an example – let’s learn something about federal spending AND get some insight into the ugly, the bad, and the good of data visualization.

I don’t usually make policy statements in this blog, but sometimes a simple and clear presentation of data says everything that needs to be said. The trend in U.S. manufacturing jobs over the last 70+ years is a great example.

For New Year’s reading I commend “What Kind of ‘Jobs President’ Has Obama Been – In 8 Charts,” from the NPR website. The eight graphs referred to are straightforward and easy to understand. They show critical data in a journalistically ethical way – a characteristic in very short supply these days. With one exception that’s all too common in data visualization today…

A couple of years ago, I wrote a series of blogs about using data visualization to appreciate the fickleness of the U.S. electorate. A group of tables and graphs examined the ups & downs in the numbers of Republicans and Democrats in the House of Representatives since 1952, compared to the political party of the incumbent president. Well, I’ve updated the graph for the 2016 election, and extended the analysis back to 1920. Read on and enjoy!

I know it’s not April Fool’s Day, when I usually write about inept data presentation, but Halloween is a reasonable second choice for this sort of thing. A graph appeared today in a Seeking Alpha blog about Gilead Sciences Inc. (GILD) that wins my black pumpkin award! HAPPY HALLOWEEN!

A fascinating article on L.A. Dodgers pitching ace Clayton Kershaw’s pursuit of an obscure but astonishing baseball record, illustrates just how powerful ratios – or key indicators, KIs, KPIs, or whatever else you want to call these metrics – can be in sports, business, personal finance, and many other walks of life. It also shows how such metrics should NOT be used.

It’s ironic that two of the most popular forms of data visualization are named after delicious but fattening and unhealthy foods: the pie chart and its developmentally challenged sibling, the doughnut chart. To illustrate my point, let’s look at how we might present the distribution of scoring on the NBA’s current top team, the Golden State Warriors. We first see a pretty but non-nutritious doughnut chart, then consider some much better ways to look at that data.

A really good example of data visualization used properly is looking at the performance of the U.S. stock market over a very long period of time. Moreover, it’s an important subject that’s critical to almost everyone’s financial security. In this post, we look at an example of data visualization that can make an intimidating subject less terrifying.

Last week, we looked at a simple, handwritten graph that’s a great example of what graphs and data visualization are really for – to help people get a sense of trends, patterns, and what behavioral consequences might be. Now let’s use that same approach to understand a typical management-by-objective (MBO) bonus plan for line managers and key employees, and improve a poorly-designed system.

Well-designed incentive compensation plans – especially sales commission plans – are an incredibly powerful way to motivate great performance. But designing a great plan is both an art and a science, and prone to design mistakes that are expensive and end up not motivating the desired performance. The commonest and most serious error plan designers make is to lay out the rules before deciding just what it is the enterprise is trying to accomplish. You can avoid that mistake with a simple, straightforward graph that I’ve drawn hundreds of times in my career.

We look at the fun subject of marital infidelity, and how it’s affected by the husband’s and the wife’s relative income. How a recent study was reported illustrates how numbers – and especially graphs – can be presented to reflect a certain point of view or agenda, often in subtle ways.

In celebration of April Fool’s Day, let’s take a look at some of the most messed-up, incomprehensible recent examples of quantation. Not surprisingly, all are graphs. But some come from sources that definitely should know better.

One of the most “boneheaded” incentive scheme design choices is excessive binary-ness. This is a problem in the Affordable Care Act, but it’s also a problem that frequently rears its head in compensation plans. In this post, we look at an example and discuss how to fix it.

We’re back, and I’m hoping you all enjoyed the holiday. As I suggested in the last post, let’s try using a table instead of a graph to see how House of Representatives election results correlate with the incumbent president’s party!

The composition of the U.S. House of Representatives provides great material for data visualization. We’ve already produced a graph (click here for the original post, here for the improved version) that, like the Minard graphic of Napoleon’s march to Moscow popularized by Edward Tufte, presents seven types of data in a single two-dimensional space. It paints a rather complete picture. But does it really feed the bulldog?

We shift gears from recent posts to address a critical – and extremely timely – decision: choosing a health insurance plan, which you must do by 12/15 if you want a new plan in place by 1/1/15. If you’re a finance professional, you will be doing your coworkers and your friends a huge favor if you can help them through this important and expensive decision. And data visualization plays a big part in this task.

I’m including this post in the series because of the extremely insightful suggestions offered by my new friend, Hicham Bou Habib, on LinkedIn’s Certified Corporate FP&A Professionals group discussion thread.

At last week’s terrific Association for Financial Professionals (AFP) national conference, there was much discussion of “data visualization,” a major information-presentation buzzword these days. Last week’s U.S. elections seem to provoke the same urges to barrage us with graphics. But do all these beautiful pictures really get us anywhere?

A fascinating piece in a recent Washington Post draws a striking comparison between income inequality and political polarization. It suggests a high correlation between greater inequality and more polarization. That’s all well and good, but it’s all meaningless if you don’t hypothesize a causal relationship between the two.

Yes, we’re still looking at Constant Contact’s (Nasdaq: CTCT) investor presentation, and the topic is still how tiny changes in the way we present numbers can have a significant impact on how our integrity and our ethics may be perceived.

Today – March 14, or 3.14 – is Pi Day, a day celebrated by mathematics aficionados the world over. I shudder to think of the partying that will break loose in two years, when the date is 3.14.16. Hide the women and children!

In recent posts we’ve seen how tiny changes in the way we present numbers can have a huge impact on how well the information is understood. In this post, we look instead at how those little things can affect how your integrity or your ethics might be perceived.

We’ve recently run posts about bad quantation in the media, about Olympic medal counts and Oscar betting odds. This time we discuss good quantation – the Food & Drug Administration’s proposed changes to the “Nutrition Facts” label. This is a great example of how small changes in the way we present numbers can have a significant impact.

It’s important to present numbers that your audience actually cares about. We sometimes lose sight of that in an effort to turn out a beautiful table or graph. I have Ira Apfel of the Association for Financial Professionals and the editor of AFP’s Exchange, to thank for a great example of this point. Here’s a snip from the “Medal Tracker” page of the ESPN website.

Happy New Year! If you’ve been reading my posts, you’ll know that if there’s one thing I dislike as much as deliberately misleading quantation, it’s pie charts. So in honor of the New Year, I want to present the most unarguably accurate and readable pie chart I’ve ever seen, sent to me courtesy of my pal Hal Ettinger.

“Chartjunk” is a term coined by Edward Tufte to describe elements in graphs or approaches to graphing that detract from the meaning of the graph or even skew the depiction of the graph to the point where it will get misinterpreted. For the interested observer, the media efforts to present information on Obamacare and its rollout have created a veritable goldmine of chartjunk.

When done properly, numbers provide a profound glimpse into the future. Here are two graphs from a Washington Post story about the possible implications of China’s recent changes to their “one-child” policy.

The Washington Post recently ran a front-page story headlined “106,000 have signed up for health coverage,” with a sub-heading of “Enrollment less than predicted.”

The law firm of Fenwick & West publishes an excellent quarterly report on venture capital financing activity. Their most recent report provides an opportunity to ponder how even modest changes in how information is presented graphically can make a big difference in how meaningful the information is to the audience.

As I said I would do last week, I now offer the first of a few posts that I hope will provide insight into the U.S. federal government’s fiscal situation, from the perspective of clear quantation. Consider the following graph (in which I break my own rule and use a pie chart) showing the distribution of federal government outlays projected for fiscal 2012 (the 12 months ending 9/30/12).

Kudos goes to Mark Whitehouse for his April 9 “Number of the Week” piece in the online Wall Street Journal. The topic is an important one – how much we spend on healthcare – and he presents some very important numbers in a clear and concise way. Here’s the chart he included in his post.

If you can understand the federal budget from what you see and hear from newscasters and soundbites from our legislators, you’re more astute than I. So take matters into your own hands: go straight to the source, and you’ll be amazed by what you can learn.

NOTE: This piece was originally posted on the 121 Silicon Valley, Inc. Website on December 31, 2010 In my last post, we introduced the idea of a “value calculator” as a powerful sales tool. To refresh your memory: The XYZ Software Company sells an add-in solution for an enterprise’s ERP system, to address situations when a single order has to be fulfilled in multiple shipments. When that happens, the enterprise’s own end customers question the invoice, causing the enterprise to have to devote man-hours to documenting and supporting a perfectly valid receivable, and delaying collections...

We all understand how critical your finance team can be in helping you negotiate and close your deals, but they can often be a huge help even earlier – in the sales process itself, by helping you make the best possible presentation of your products’ value proposition.