Text

Data Visualization Fundamentals course on Lynda.com

image

After many many months of writing, planning, gathering examples, designing curriculum, preparing materials and recording, I’m thrilled to announce the launch of my course on Lynda.com, titled Data Visualization Fundamentals.

In this course, beginners to the field of data visualization will learn a broad range about why visual communications is so important, including some of the brain science behind this, and how to take intangible data and make it relatable, visual and approachable for your audience. With over 3 1/2 hours of content, the lessons include quick studies of topics like:

—Channeling your audience
—Understanding your data
—Determining the information hierarchy
—Sketching and wireframing your ideas
—Defining your narrative
—Using typography, color, contrast, and shape to convey meaning
—Making your visualization interactive

Check it out and let me know what you think! You can see the welcome video below.

Text

Does Your Company Actually Need Data Visualization?

image

The Harvard Business Review published an article of mine titled "Does Your Company Actually Need Data Visualization?”. There I talk about who should (and shouldn’t) bother with data viz. Take a read and feel free to comment there or email me or comment on my Tumblr about it. I think that many organizations should be visualizing (and making their data) interactive, but not all. 

http://blogs.hbr.org/2013/11/does-your-company-actually-need-data-visualization/

I look forward to hearing what you think!

Text

Journalism = Teaching = Information Design

image

I’m giving a workshop in a couple of weeks on “Thinking Visually and Interactively” for the Association of Management Consulting Firms. I’ve also been approached to teach an online course on the Foundations of Information Design. It’s an honor to be in a position to share my thoughts on these subjects, which are my passion.

Teaching is such a special and unique way of thinking about and communicating a subject. Especially when the audience is assumed to be mostly novices. It forces me to remember what it was like to approach this work before I started doing what I’m doing. It takes me back almost 20 years. And that is a fun and interesting place to visit. 

Twenty years ago, the technology was very different. Broadband didn’t exist. (I probably had a 14,400 baud dial-up modem on my computer!) The web hadn’t come close to the mainstream yet. But information design was, of course, alive and well. Magazines, newspapers, television - all the visual media of the time were presenting information in ways meant to help their audiences understand a topic. So in the most important ways, nothing has changed. Oh, and 20 years ago I was a journalist.

When I started doing web design and development, I thought I had abandoned journalism to a large degree. But the longer I worked, the more I realized that I was leaning heavily on that knowledge and experience.

And now, with teaching, even more so. Journalism is almost the same thing as teaching. You have a set of complex information. It needs to be organized and a story needs to be told. This is about giving hierarchy to the information and explaining it in a way that the least informed recipient of that information can understand it, while the more informed can find nuances and details that will satisfy.

Journalism and teaching are both, first and foremost, about creating a “curriculum” based on the information at hand, designing a presentation layer for that information (in words, lesson plans, visuals, etc.) and presenting that information to increase knowledge. The outcomes are a bit different but not by much.

And, by the way, the description above also captures one definition of information design. So really, Journalism = Teaching = Information Design. I could probably keep writing and eventually get to the point where I talk myself into all three of those being equal to sea kelp and solar panels too, but in all seriousness, there’s something to this…

What do you think? What are the differences between these practices that I’m missing?

Text

Hospital Pricing Visualization Challenge

I recently submitted an entry to a data visualization contest sponsored by the Robert Wood Johnson Foundation, Health 2.0 and Visualizing.org. The challenge was to create a static or interactive visualization of data on hospital pricing provided by Centers for Medicare and Medicaid Services. I’ll find out in mid-September whether I won, which could mean some cash and, more importantly, an opportunity to present at the Health 2.0 conference.

My entry takes the point of view of a consumer looking for care for him/herself. I wanted a tool that would make it easy to see relevant data to help me find medical care based on quality and pricing. 

What’s nice is that I’ve already had some good market validation for the visualization. It was picked up by Reuters and Popular Science last week, which has led to a good amount of traffic to the site! It was also selected as a Featured Visualization by Visualizing.org and when I cross-posted it to visual.ly, it was chosen as a Staff Pick there! 

Text

Thought Leadership, Your Pipeline and Visualization

image

Your Pipeline Has a Leak

95% of your potential sales pipeline (the universe of people that COULD buy your services) is going to produce zip - not even an opportunity to pitch your services. Is there anything you can and should be doing to maybe make that 94% or 93% or less? Can you bump up your sales opportunities by concentrating on your marketing and sales to increase those odds? Of course you can.

Can interactive data visualization play a key role in this process? I believe it can.

To sell, you need to show a prospect that you know them, you understand their business, their industry, their particular challenges. You’re probably great at doing that once you’ve had a few calls with them and are in the room for the big pitch for their business. But how can you do that before you’ve met them, before you’ve exchanged an email with them, before you’re even aware that they exist?

Marketing Plugs Leaks

This is exactly what marketing is supposed to do, really. Marketing is showing the market that you exist, and that your solutions and ideas fit with the world’s needs. And thought leadership, in particular, shows the world that you have specific smart ideas that can help solve real problems. It keeps you on the list for a prospect - it keeps the pipeline from leaking.

But your thought leadership is editorialized. You’ve done your research and teased out your conclusions and are publishing a specific analysis and storyline. Sometimes, especially for research- and data-driven thought leadership, you are necessarily leaving something (often a lot of things) out. You are not telling your audience’s story, you’re not connecting with each reader, but rather a homogenized median reader. An average; a profile. 

Interactive Delivers Granularity and Personalization Your Audience Needs

Interactive data visualizations solve this. By opening up the kimono and sharing your data and the tools to allow your audience to play with the data themselves, you are empowering them to see themselves in your data. Perhaps you’re letting them find how the data relates specifically to companies in their industry or their region or their size. Either way (and 100 ways more), you’re doing what I proposed marketing is supposed to to do: you’re showing the market that you exist and that you understand your audience. You really understand them - your data shows that your ideas apply to them directly, not some amalgam of companies in a certain category sort of like them.

Data visualization does something else that no PDF will ever do. It attracts the attention of a broader audience and encourages sharing. Not too many people are trolling the internet for another 25 page PDF. But people are trolling for engaging and interesting visual and interactive experiences.  And they will share the most engaging ones with friends and colleagues. So while the CEO of the Fortune 1000 firm may never stumble across your interactive visualization on his or her own, you can bet that the junior associate six levels down will find it, will share it with colleagues and managers, and if it’s great and really applies to their business, it will trickle up to the CEO.

If you could turn 1% more of the universe of potential business into an actionable lead, what would that mean to your organization?

Photo credit: http://www.flickr.com/photos/willowherb/

Text

The Right Visualization for the Job

Some people seem to think that data visualization has to be extreme, fancy, complicated and dense. I’ve seen visualizations that are incredibly beautiful and awe-inspiring as a design or programming portfolio piece. But they do little to illuminate the subject at hand.

The other end of the spectrum is the lowly bar chart or pie chart. These basic forms of data visualization are, in fact, useful and should never be ignored. In fact, sometimes they are the right visualization for the job.

I just created a new visualization to illustrate the data from a survey on how consulting firms are using social media. Conducted by the Association of Management Consulting Firms, The Bloom Group and ResearchNow, it reveals a lot of interesting data and trends in the industry’s use of social.

image

Perhaps the most interesting is the difference between the Leaders and Laggards. Leaders are those whose thought leadership programs generate more than 30 leads per month and Laggards generate fewer than ten per month. Without getting into the research itself (you can read about it on The Bloom Group’s site and see the visualization here), I want to explain my reasoning for the approach to the visualization.

Leaders vs Laggards

My first question is always “what’s the story”? And the story here is definitely Leaders vs. Laggards. So the key thing the visualization had to achieve was to display the gap between the two. A simple bar chart could have displayed this – the difference between the heights of the bars is clear. However, the difference isn’t as visible for two bars when one is 6% and the other is 3% unless you set the y scale to a low number, which is a bit disingenuous – especially in a format people are used to seeing in a certain way.

So I decided to create horizontal bars that originate on the outside edges and animate inward, on a proportional scale (so whether it’s 6% vs. 3% or 60% vs. 30%, the bars always meet in the middle.) Because it’s a non-traditional format, I think people will forgive the scale adjustment. And the numbers are clearly visible so you can see that you’re not looking at a 100% scale. The gaps are highlighted to reveal the difference between the two. This makes it obvious that 6% vs 3% is a 100% variance, and 36% vs. 33% is a much closer gap, even though in both cases we’re talking about 3%.

To make it really clear that this is not some attempt to over-emphasize the differences while discounting the real percentages, you can also view the data using a more traditional 100% scale view. In this case, I made the bars originate in the center and animate outward so you can see them next to each other and more easily see the difference in size between the two.

Compare What You Want to Compare

The other key feature of this visualization is the ability to view more than one set of bars at a time. Rather than pre-package the data in groupings (showing all the 2013 data or all the data for a specific question, for example), the user can choose which sets of data s/he wants to compare and can look at as many at a time as s/he wants. This makes it a very powerful comparison tool while reducing restrictions placed on the user.

Simple, traditional shapes, nothing fancy really, but a powerful tool to see and clearly understand the data at hand. Play with it yourself and let me know what you think!

Text

To Interact or Not to Interact?

image

The world is getting more visual every day. See the meteoric rise of Instagram and Pinterest and the explosion of infographics as recent proof points. People want to see information, not simply read about it. This is, in part, due to the ever-expanding amount of information we are processing on a daily basis. We have trained ourselves to scan things quickly, seeking nuggets of interest upon which we will focus our limited time. If you are presenting complicated information that includes data, you must make it visual to transform your audience from slightly interested bystanders into active consumers of your content.

Every two days, humankind generates more data than was created since the dawn of time through 2003, according to Eric Schmidt. Think about that for a second - from civilization 0.1 through 2003…we’re not talking about 1900 or even 1990. This is a staggering statistic. And that quote is from mid-2010 - I’m almost afraid to hear it updated now, 2 1/2 years later.

Data is abundant (understatement!) and to understand it - for it to become knowledge, not simply information - requires we can clearly and simply analyze and reflect on it. The old way of presenting this information isn’t good enough. We need new ways of communicating data to allow our audiences to truly learn and make decisions from it, including whether to hire you for that consulting gig or to do additional research, to buy your product, etc.

Data Tangibility (Interactivity) Works

Infographics, which seem to have become ubiquitous in 2012, are a good start, but aren’t enough in many cases. And the question facing researchers, thought leadership marketers and communications professionals generally, is when is a visual display of data enough and when do you need to take it to the next level and make it truly tangible? When and how do you turn those formerly passive bystanders who are now interested consumers of your content into active participants in it? There are two primary ways to think about why going interactive makes sense:

  1. Cognitive Fit Theory is the result of an academic paper that, in a nutshell, proved that to help people solve problems, it is critical to understand the nature of the task itself as well as the problem-solving skills of the participant and to present the challenge in such a way so that it maximizes the match between them. In other words, if you want your audience to learn something, take an action, change their mind, you need to present your data in a way that takes advantage of what you know about how they think and interact with your information. Iris Vessey, who co-authored this study, did more research that found that adding interactivity to information presentation leads to better decision-making from participants. Simply put: making it interactive makes it more impactful and powerful.
  2. People are, to varying degrees, narcissistic. I don’t say that to be flip. It’s just that we naturally tend to ask “so what - what’s that got to do with me?” Hearing a story about someone getting eaten by a tiger is nothing compared to the visceral reaction upon seeing a tiger bearing down on you from 30 feet away. When you make your data interactive, you are empowering your audience to make the data about them. Rather than limiting your audience based on the editorial decisions made when preparing a thought leadership article or even an infographic, an interactive data experience allows them to click and filter and select the data stories most relevant to them. By empowering them with open access, you are allowing them to use your data to tell their own stories, in the context of the argument you’re presenting.

Examples

One great example of how empowering interactive data can be is from the New York Times, which creates many such experiences. In 2011, the United States was facing a loud and boisterous debate about our national finances. (Sound familiar?) The Times created an interactive titled “Budget Puzzle: You Fix the Budget”, which allowed users to make choices - from cutting earmarks or defense spending, to increasing the eligibility age for Medicare and Social Security and raising taxes on the wealthy. All in all, there were 20 choices. As you make selections, you see the budget imbalances cured (or not.) More effective than a pundit blaring the party line at you, the tool makes it immediately clear that this is a debate with impossibly difficult decisions and no easy outs. You can make it about you and your own priorities - see for yourself.

Booz & Company came out with an interesting report last year about M&A showing that deals done for capabilities fit (as opposed to market share grab or geographic reach, for instance), return 12% better results to shareholders than those that don’t. It was an interesting study but the interactive piece we created for them revealed much more detail than could be covered in the Strategy + Business article. For instance, by clicking through the data, users could see some important details - such as how this research applies to their industry, not just those highlighted in the study. IT, for example, clearly reflects the overall trend, whereas the data was less aligned for Electric Utilities. The interactive also revealed some things the report wasn’t intended to cover - such as the overall terrible performance of M&A in the Media sector, and how Healthcare deals seem to be done almost exclusively for capabilities fit. And you could look at all 320 individual deals to see how an individual company appears in the study, which I regularly do when meeting with companies who have done M&A in recent years.

Not only does using interactive data experiences create better outcomes than static (even visual) data, it also drives better web analytics. We’ve worked for many years on interactive experiences and regularly see much better performance on those projects than their static thought leadership article counterparts. Bounce rates and exit rates (the percentage of people who leave your site after viewing a particular web page) go down significantly (25-50%), social sharing goes up (usually at least doubling the number of shares) and even the long tail of visits looks much better. An article will have a spike and then traffic falls off quickly and somewhat permanently, whereas interactive pieces have ongoing traffic for a longer period of time - likely due to the amount of social sharing.

To Interact or Not To Interact?

It can be daunting to decide whether to make your data interactive. Budget and time investments need to be weighed against the upside potential. The biggest barrier tends to be inertia - the “we’ve always done it this way, so why change?” argument. But keep in mind that budgets don’t have to be big, and if you’re focused and smart about which pieces you make interactive, the benefits are clear and compelling. Interactive tools also have an advantage over their static counterparts in that investments made can be often repurposed for pennies on the dollar for other data experiences. Reusability and phased development are a smart way to test the waters and build tools that can have a long shelf life. As data continues to grow exponentially and your audience’s attention grows ever more scarce, you have to be visual to break through the clutter. And once you have them, take a targeted approach to making your data tangible and your audience will be more likely to learn and act based on the real knowledge you expose to them.

Text

The Best Hockey Team Ever

So it turns out the best hockey team ever…was the 1929-1930 Boston Bruins. Yep. That’s right.

image

Data visualization is great. It lets you take the numbers, play with them for awhile, and find a way to make sure that the Boston Bruins come out on top!

But seriously, there are a lot of interesting stories in here. (Play with the interactive version to see for yourself.) For instance, look at the lowly 1974-1975 Washington Capitals. Ouch! And what happened in 1992-1993? There are two different teams that only won 14% of the possible points that season. Sorry Ottawa and San Jose, that must have really hurt.

And notice the Montreal Canadiens who, I must admit, hold a lot of those top spots with a bunch of Bruins teams. Of course, the Canadiens did take those 23 Stanley Cups compared to Boston’s…can’t remember how many…but it’s less than that…hmmmmmm. Still, the data proves that the Boston Bruins are the Best Team EVAAARRRRRR!

Note, this was not designed to run in Internet Explorer. In fact, it reminded me (this happens frequently) why I hate IE so much. The other browsers are so so much better. This was built in D3.

Text

Interactive = Mass Customization (which your clients want)

"Enough about me. Let’s talk about you…What do you think about me?"

I love that joke. It is the perfect metaphor for so many things. And it’s something that I refer to frequently when talking to clients.

image

The fact is, people are obsessed with themselves. OK, maybe that’s a bit strong. But it is fair to say that if you can make what you’re communicating about me, and not just about some random “other” who may or may not be like me, I am far more likely to listen to you, really hear you, understand you and, perhaps most importantly, remember you (and your message.)

Source For Consulting recently released its annual study of clients of consulting companies to get their insights and opinions into why they use consulting services. They decided to ask some questions specifically about thought leadership marketing as well. And there were some really interesting findings in there. One in particular stood out to me.

Relevance = Personalization

Clients want thought leadership content to be relevant to them. In fact, the most important thing they identified as being an “essential” quality of thought leadership they consume, is its “relevance to your industry/function/organization”.

image

As Source for Consulting concluded, they essentially want “mass customization”. What does this mean? How can you leverage the work you’re already doing to provide quality thought leadership to your audience – to make it about them – if your editorial process needs to narrow in on messages that may not include them? For example, if your research returns data that is particularly interesting for the Energy sector, odds are you’ll focus on that portion of the data to illustrate your point. But that doesn’t mean a Healthcare executive might not find a compelling story in your data.

Can you afford to leave him or her out? On the other hand, can you afford to devote the time and resources to develop a report that slices and dices, analyzes and summarizes the data for every possible audience and angle? Of course not, but this is one of the best examples of how an interactive data experience can help.

image

Expose Yourself…or at least your data

You need to expose your data. Put up an interactive graphic allowing your users to play with the data and slice and dice by industry, by company size, by geography…by whatever criteria that will make it relevant to them. You don’t have to make a massive investment in some complicated platform. Sometimes simply the ability to see some basic charts from different angles is enough. Other times your thought leadership may warrant a rich, broad and deep data exploration platform. (Maybe you’re launching your biggest initiative of the year at Davos and it’s time to pull out all the stops.)

The point is that you should do everything you can to empower your audience with information, make it as much about them as possible, and give them knowledge on which they can act. If you make it all about them, and they can see a story where previously there was none for them, you’ve created an opportunity for a connection, a phone call, an outcome that would have otherwise been missed.

Can you share any examples of your favorite interactive data experiences that let you make someone’s data relevant to you?

Text

ASK… What Makes a Good Data Visualization

There’s good data visualization and there’s bad data visualization. Rather than quote Supreme Court Justice Potter Stewart’s definition of porn (read Jacobellis v. Ohio), I think there is actually a very specific answer to the question what makes a good data visualization? It can be summarized in the acronym ASK.

A = Accuracy

First, a good visualization must be Accurate. I love to criticize Fox News’ many attempts to skew its charts and graphs to match its political agenda. One of the more interesting examples of this is from 2011 when they showed a chart implying that the unemployment rate under Obama was stagnant or rising when, in fact, it was declining.

image

As Media Matters correctly pointed out, there were plenty of problems with this graphic. Fox News shows an “alarming pattern of dishonesty”, as they say. Specifically, their graph seems to intentionally put the dots on the graph in the wrong place to make lower numbers seem higher.

image

I would also argue that the scale was off (the y axis should have gone from 8.6 to 9.2 or maybe 8.5-9.5 at most, rather than 8-10, which over-emphasizes the static nature of the numbers.) The chart below is the one originally provided by the Bureau of Labor Statistics.

image

Accuracy is the backbone and a baseline assumption of any good visualization.

S = Story

Second, a good visualization tells a story. Edward Tufte (widely considered the father of modern data visualization) drew attention to Charles Joseph Minard’s famous visualization of Napoelon’s March as a quintessential example of visualization.

image

And once you’ve seen it, it is very unlikely that you will forget it. This is really due to the story it tells. In its simplest form, the story you can see is that as Napoleon led his army to Moscow in the winter of 1812-1813, he lost 300,000 men (75% of his army). When he decided to retreat, he lost most of those men who had survived the initial march. As Étienne-Jules Marey said, this graphic “defies the pen of the historian in its brutal eloquence”. That’s story-telling.

K = Knowledge

It seems obvious that a visualization should impart knowledge. Rather than a pile of papers or a thousand rows in Excel, a chart boils down the data to its essence, with the story visible at a glance. But it’s not always true that a visualization provides knowledge even if it is accurate and is telling a story.

Consider the New York Magazine graphic explaining four decades of the trials and tribulations of All My Children’s 200+ characters. I’ll take it as a given that it’s accurate. And it certainly tells a story (I won’t judge the show by commenting on it) but whooee, it does not impart knowledge. It’s far too dense with information for me to really learn anything from it. In fact, it makes me want to nap.

ASK Yourself

So remember – the next time you have data to share and you’re thinking about visualizing it (which is a great idea, by the way), I recommend that you ASK yourself what makes a good data visualization. And act on that knowledge.