Sticky slides – which is the best chart?

One of the many RSS subscriptions that keep my Google Reader working overtime is Jan Schultink’s Sticky Slides blog, which provides a lot of useful advice around presentation design. Today Jan has published a post comparing two alternative chart designs that present the same data from a survey question.

Jan’s view is that the second graphic is the more effective:

Both of these charts contain the exact same data. The second is a lot easier to read, the spectrum of customer choices is neatly laid out, and the colors are picked in sequential order.

My view is that the first graphic is, overall, the more effective. It does contain unnecessary value labels, I’d also remove the question reference codes and there is an execution blemish on the Japan chart compared to the others (presumably caused by an inconsistent paste special selection), but I think it is a much more elegant and principles-compliant approach that allows readers to draw insight from the distribution of buying attitudes in each country.

What do you think?


Joe MakoSeptember 10th, 2010 at 9:18 am

Here is how I would viz it:

Andy KirkSeptember 10th, 2010 at 9:42 am

Wow fast work Joe! Thanks for sharing that -it does work well in Tableau especially with its interactivity for filtering the individual countries. I’d be tempted to switch off the data labels and there is the issue of whether the sample sizes for each country were the same or not, but of course we don’t know this so as an illustration of an approach I like the inclusion of average.

Andrew MarrittSeptember 10th, 2010 at 12:03 pm

I haven’t bothered to tidy this up or taken much care on formatting.

I think the information that you really want to show with this type of question is the amount of people that would buy against the number of people who wouldn’t. In some ways the ‘maybe’ might not be that interested (like a neither agree or disagree response). For me it’s the strengths of feelings.

I wanted to center staked bar on the centre of the ‘maybe’ but with only 10 days Tableau experience couldn’t work that one out (anyone who knows how to do it please let me know). I therefore excluded it and changed the negative views to negative numbers. You get the following:

This view, I believe, does the best job of showing the positive and negative feelings about the product.

Jan SchultinkSeptember 10th, 2010 at 2:49 pm

Thank you for taking the discussion further here.

Andy KirkSeptember 10th, 2010 at 5:48 pm

Thanks for the contribution Andrew that’s an interesting alternative approach, good luck with developing tour tableau knowledge!

Emeka NwokoyeSeptember 13th, 2010 at 10:36 pm

Hi guys,
Quick question, please could you tell me what software or design package you used to construct your data charts? I’m at university at the moment and I’m planning on focusing on information design for one of my assignments. I just need some insight as to the best programs to use to display information such as the ones above.

Many thanks guys!

Emeka Nwokoye

Andy KirkSeptember 14th, 2010 at 7:47 am

Hi Emeka, thanks for your message. I would imagine the original charts were produced using Excel. The guys who have posted alternatives within the comments have used ‘Tableau public’ which is a very powerful, very important visualisation tool and importantly free to use. If you can wait until October, I should finally get round to publishing a series of posts which present a massive range of visualisation resources both practical and theoertical (and both free and paid for). I was hoping to have completed it by now but it has been a difficult list to tame as there is so much! All the best with your assignment Emeka

Emeka NwokoyeSeptember 27th, 2010 at 10:55 am

Very much appreciated Andy! I will investigate Tableau Public!

Thank you very much!