Comments about oreilly Information Dashboard Design:

This book changed my life when I first bought it. I still re-read it every few months and use its ideas frequently in my work.

Chapter 4 alone is worth the full purchase price of the book (various visual techniques distinguished by their usefulness for communicating numeric values).

I've implemented many of the ideas in my own personal reporting tools and have found them to be extremely useful.

I wish the developers of the many services I use would read this book so they can go beyond pretty graphics toward making data easier to understand and interpret through visual presentation.

Comments about oreilly Information Dashboard Design:

The materials are quite basic, consists mainly of my chapter 2 from my Freshman Statistics For Managers book (Prentice Hall). Nothing new. Save your money.

Comments about oreilly Information Dashboard Design:

The reason I wanted to get a copy of this book to review was that I was hugely frustrated by the volume of data I had to scan for patterns at a previous job. Now I have moved on I have had time to revisit the book and do a better review.

Description:

Information Dashboard Design is a book which can educate the reader so that they can design more effective information dashboards. The book defines a dashboard thus:

"A dashboard is a visual display of the most important information needed to achieve one or more objectives which fits entirely on a single computer screen so it can be monitored at a glance."

In about two hundred large, uncluttered, and well illustrated pages Stephen Few takes the reader through: A historical survey to help define dashboards and put them in context, thirteen commonly made mistakes, a quick overview of human visual perception, mixing simplicity and effective media to make a usable dashboard, and then some case studies to tie it all together with real examples.

What's to like?

The book appeals to me in its large format with uncluttered pages. By the time I reached the end of the book I appreciated that Stephen had used many of his techniques for simplifying and removing clutter information dashboards to make the book more effective. As a result the book struck me as almost a coffee table book, beautiful enough just to leaf through for its own sake.

The pacing of the presentation meant that there was enough mental space for me to digest the points being made.

The structure is logical and flows well, Stephen's notion of effective presentation of data pretty much coincides with Edward Tufte's (a good thing, in my opinion), and the examples are relevant and useful. I particularly liked the critiques of designs which could be improved - many times I have had the intuition that a dashboard is not so good, and the criticism of the example dashboards helped me understand why they were not as good as they could be.

The coverage of human perception was interesting to me (even though I have a red/green colour blind brother I had never considered what he might miss in a dashboard using traffic light style colours), and this overview gave an underpinning to the design of some media which could be read at a glance even though they were information dense.

The final set of examples pulled all of the advice together, and it was a delight to compare and contrast the "good" dashboards at the end of the book with the "bad" examples presented earlier. As each of the final examples had a different prospective user I could see how the advice and guidance can be used for different users, and that they met the author's criteria for well designed dashboards:

"Well-designed dashboards deliver information that is:

* Exceptionally well organized

* Condensed, primarily in the form of summaries and exceptions

* Specific to and customized for the dashboard's audience and objectives

* Displayed using concise and often small media that communicate the data and its message in the clearest and most direct way possible"

What's not to like?

Overall I liked the book, and only occasionally did the tone seem to become a little "distant". I would have liked more examples, and I suspect that that's more my laziness than the author's oversight.

Conclusion

I liked the book, and I think that it will be useful for me many times in the future. The techniques used to distill and condense data for use in dashboards are useful in many other contexts, and the information about perception has got me thinking about all sorts of other places where I can use our perceptual mechanisms to make users' lives easier.

The book is useful if you are designing dashboards for others, or even generating email reports for yourself which need an overview section which can be interpreted at a glance.

Comments about oreilly Information Dashboard Design:

I like that the material is there and it's useful: dashboard design mistakes to avoid, useful display guidelines, theory, etc.

I like that we both seem to think along the same lines: keep the display simple, functional, and pertinent.

There is one thing that I do not like: Many of the captions are along the lines of "This dashboard has problems." I would rather have seen the problems explicitly mentioned, possibly tying it together with the text a bit better.

And there is one thing I am at odds with: Treemaps. Stephen seems to love them (for the "right" data), but the example provided just seems to be difficult to decipher.

I've recently used the book as a guide for a dashboard created for a friend, and the result is anything but fancy, but it is very functional and useful. It is web-based, and has a single update button. Some of the data is updated on demand, while other data is updated on an appropriate schedule and drawn from temporary locations for display. At the top are a few key numbers, which are highlighted with red, yellow, or green backgrounds to indicate potentially severe issues that should be checked immediately, numbers that should be monitored closely, or numbers that are in the desired range, respectively. There is another group of data below that, showing numbers, percentages, and a horizontal bars which give relationships at a glance. A similar group showing the same information for an extended time period is below that (and sometimes is below the screen display, depending on the browser size or the browser text size - Stephen says that this is a no-no, since a dashboard is meant to display all key information at a glance; however, it is more of a convenience, since none of the other data need to have drill downs, and there's no reason to lose the main information just to show the secondary information. It is a simple dashboard, and a very minor exception).

I am disappointed that I was unable to make use of Stephen's "Bullet Graph," a very concise graph tool that displays a key measure, along with a comparative measure, and qualitative ranges. It looks like a bar graph, but is much more sophisticated. You'll need to get the book to see it in action.

It is a useful book. I much prefer the negative and positive guidelines, to the theory, but it is all there.

Review originally posted at: http://elug.ca/reviews/information_dashboard_design.shtml