http://www.juiceanalytics.com/writing/core-knowledge-pack-for-data-analysts/

My approach to visualisation is based upon flying and particularly landing a helicopter on an oil rig out at sea. Firstly, a ‘Red alert’, management needs to take immediate action, ‘Amber alert’ be aware that corrective action is needed to correct an aspect of the process, and finally trend monitoring – a quick scan to monitor process performance. Similar to flying your eyes need to be outside of the cockpit the majority of the time, with only a quick scan to monitor that all systems are OK. If management cannot understand my dashboard within 20 seconds then I have failed.

So to answer the exam question core skills: are you looking for a doctor or a nurse?
If the doctor then sound business skills, a bag full of business and analysis tools, good bedside manner and ability to listen. And armed with Stephen Few’s methodology on dash board design you are ready to go!

Given our limited resources (time), the challenge is to know what we need to know and prioritize it accordingly. Do we need a good default/template? Do we need to become a specialist? What about learning the key points just to remove some roadblocks somewhere else? There is no black and white.

(Great post, by the way. Thanks for sharing.)

In this case, knowing is understanding the data, learning what matters.

Planning is when you draw some rough sketches of the dashboards (or charts or reports), decide on alternatives and figure out the final layout.

Implementation is when the tool and techniques come in to picture. In my case Excel is the tool and techniques are knowing the formulas, data massaging techniques and obviously charting (and a bit of trickery like conditional formatting, vba, form controls etc.)

Here is a post where all three are shown to make a project status dashboard: http://chandoo.org/wp/2009/10/06/project-status-dashboard/

While I agree with your “teach fishing” theory, most of the time people dont want to become fishers. They want to become hunters, dancers or coffee guzzlers. They just want fish for a side dish

I think a moderate approach like give the fish, but encourage readers to learn fishing is better. That is what I do on PHD. From time to time I provide free templates (or even paid templates), but also write tutorials on how to do these. It has been working alright and readers seem to enjoy both.

Jon: You are factually right. Kyd is emphasizing his ready-made dashboards (like the IncSight QnE) and I was forgetting about the e-book – I actually bought it a few years ago. I’m using his example because he targets those users that are more interested on a fast implementation with their current skills than on getting new skills (well, that’s my perception, anyway…).

Actually, Excel User started with an ebook showing how to build robust updatable dashboards. This expanded with some example workbooks to show how to implement the techniques in the book. When customers began clamoring for workbooks that were more extensive than the examples and faster to implement, Excel User listened to the customers and came out with their ready-made dashboards.

So in addition to the fish, Excel User provides the know-how to fish, for those so inclined.

The task is the business need. Don’t stop with the specifications that your manager gave you or past practice. Dig deeper and find out how the chart will be (or might be) used.

To this task, apply the appropriate technique. Here, there is no substitute for some academic training in statistics. Anyone can plot a time series, but what is the best way to extrapolate a particular trend? When should you use a scatterplot, and what kind of trendline should you have? Might a quantile function be a good way to present a risk analysis? Some old-school background would come in handy on these questions.

Finally, understand the tool–in this case, Excel. All the theory in the world won’t matter if you can’t put it on the page.