There are only two kinds of data visualization in the modern world. They are Story Visualizations and Answer Visualizations.
Story Visualizations are those produced by one set of people with the goal of telling a story to an audience. Think of a newspaper graph showing deaths during a war, or a map showing where within the country unemployment is highest.
The second kind of visualisations are Answer Visualisations. Answer Visualizations are produced to supply an answer to a single question posed by a particular person. Before computers, answer vizualisations were slow to produce and the preserve of narrow elites: a CEO asking for a graph of product sales, or an engineer constructing a graph for their own needs. In the Internet age answer visualisation has become so quick and common that sometimes you don't even notice their presence. Think of the unobstrusive maps that pop up in Google search results when you type in a business name.
I think that the failure to understand the difference between Story Vizualisations and Answer Vizualisations is at the root cause of a lot of people's confusion and frustration with the state of contemporary data visualisation in general.
The confusion comes particularly strongly to people who commission visualization projects which they expect to be amazing only to feel strangely underwhelmed at the swishing, zooming deliverable that shows up. This is because they have been expecting something that will deliver the sense of wonder and satisfaction that comes from familiar interactive experiences like games or social media. But a Story Vizualisation can never do that because it doesn't put the individual user at the heart of the experience: it's not about you, you're just an audience member, there to gawp.
This isn't for a moment to scorn the importance of Story Visualisations, which are a vital and often beautiful part of human communication. But Story Visualisations can never be more interesting than the story. And it is in this regard that so many shiny-shiny Story Visualisations fall down: the average stories in the shiny packaging are like a VW Beatle engine in a Ferrari's body. The whole thing disappoints much more than seeing either a Beatle or a Ferrari would on their own.
Answer Visualisations are inately more about you, and what you want. This appeals to our narcissism, and our gratitude at being helped. But they're much harder to build than Story Visualisations: you can't make them in Excel or Google Docs. In fact, they're almost entirely the preserve of the coder-classes at the moment.
And this is where the frustration comes from. People can touch and taste Answer Visualisations every day, but they cannot make them. Not only that, but when they ask experts to help, they can be shocked or surprised by the costs, and confused by the imponderable questions thrown up by designing dynamic systems. "Isn't this stuff all easy and free now? I saw my daugher make a lovely graph for school in five minutes". They feel, rightly, that this shouldn't be so expensive or confusing.
And that is my last thought. With Excel and Socrata and Google Docs and ManyEyes we do really all have what it takes to make all but the most fancy Story Visualisations ourselves. But lowering the barriers to Answer Visualisations seems almost unimaginably hard and distant. The best we can do for now is understand and explain the difference, and contribute to efforts to bring Answer Vizualisations closer to the reach of everyday users.
End Note: There is actually a third kind of visualisation, the kind which contains many answers to many questions, all compressed into one graphic. A normal map or a sine/cosine table are examples. But with a few exceptions these are mostly being made obsolete by tools that produce answer vizualisations.