Discussion


Note: Images may be inserted into your messages by uploading a file attachment (see "Manage Attachments"). Even though it doesn't appear when previewed, the image will appear at the end of your message once it is posted.
Register Latest Topics
 
 
 


Reply
  Author   Comment  
dnguyen

Registered:
Posts: 21
Reply with quote  #1 
I thought the attached charts deserve to be posted here :)

1. "Tree & Shrub Sale" flier created by a non-profit org based in Maryland. I think it's a clever use of the "horizontal bar graph"

2. "Mini square" as alternative to pie chart extracted from Smart Growth poster (created by National Geographic back in 2002): I love it and definitely effective for a poster that's intended for the general public. If you know the official name for this chart, please let me know.

3. "Accessibility to Commuter Services" pie chart extracted from Smart Growth poster (created by Maryland Department of Planning in 2010). Ignoring the clutters (font placement, colors) it's a "good pie."

Attached Images
Name: HorBarChart.png, Views: 1899, Size: 794.02 KB

Name: PieAlternative.png, Views: 1898, Size: 108.77 KB

Name: anotherPie.png, Views: 1895, Size: 64.79 KB


sfew

Moderator
Registered:
Posts: 812
Reply with quote  #2 
Duc,

Do you really think that these charts are effectively designed? In all three cases a simple bar graph would present the information much more effectively. Each of these charts displays a different problem in the representation of quantitative information: the pie chart forces us to compare areas and angles, which human visual perception cannot do effectively; we have no idea how to compare the trees (can you put them in order by value from highest to lowest?); the squares must be counted to compare the values.

__________________
Stephen Few
dnguyen

Registered:
Posts: 21
Reply with quote  #3 
Steve,

I do not think the charts were designed to their purest form looking from a quantitative view point (especially the pie chart). However, they probably work well from a qualitative view point.

For the "tree and shrub sale" flier, I think the creator wants to give the audience a sense of 'scale' as in how big it may become ... well assuming all factor influence the growth are held constant. Given that it's a flier, I think it's probably tolerable (well for me that is :) )

For the mini square, when I read it I get a sense of absolute space in relation to the amount of total land that can be managed in Maryland. Each category may change (depending on the changing policies) in the expense of the other categories. This type of message is important I guess in some form of communication even though it's not the most pure method to represent quantitative information. What do others readers think?
dnguyen

Registered:
Posts: 21
Reply with quote  #4 
Steve and others,
I found a post on Juice Analytics (dated 2006). They challenged folks to create alternative to pie charts in excel and picked the "square-pie" version to the "best alternative"
http://www.juiceanalytics.com/writing/solving-the-pie/
I love to hear your thoughts in regards to this approach. Should people (including me) start using this type of visuals or just ignore it completely? If the recommendation is maybe, then under what circumstances do they work best?
Thanks.

wd

Registered:
Posts: 167
Reply with quote  #5 
Duc;

Take a look (again - you've been there before) at Cleveland's Graphical Hierarchy elsewhere in the discussion area.  Square pies use area to represent the data -  one of the weakest methods to convey information.

__________________
Bill Droogendyk
sfew

Moderator
Registered:
Posts: 812
Reply with quote  #6 
Duc,

A "Square Pie", or whatever you choose to call this graphic, is a popular alternative to pie charts in many journalistic infographics, but not an effective alternative. To compare the parts, you must either compare areas, which visual perception doesn't handle well, or count the squares, which is slow. A normal bar graph makes the comparison of parts easy and precise.

__________________
Stephen Few
Anders

Registered:
Posts: 18
Reply with quote  #7 
I can't believe anybody who've read this blog would post a pie chart in this section...
All three graphics are pretty weak. However, rotate the "Tree & Shrub Sale" graphic and assume the scale is correct, it could be used to compare typical height, width, shape and leave thickness, etc. for the trees and bushes. Maybe it's even possible to identify the trees just by this graphic. In that case, it does encode more information then a bar chart, but the bar chart would still be better for quantitative data like height.

The square chart is just stupid; just put in a number instead. The square charts in the link to juice analytics is even worse; filling the chart from the corner and making it even harder to compare the areas.
jlbriggs

Registered:
Posts: 194
Reply with quote  #8 
I don't think it's fair to call the shrub and tree sale flier a weak graphic, given its context: it's a sale flier.  it's not a chart in any real sense, and I'm certain that the purpose behind it was far more on the side of entertainment than data communication.

As a chart, sure, it's ineffective.  But it's not meant to be a chart...

Don't mean to derail anything, just felt that was important to distinguish.  :)

dland

Registered:
Posts: 1
Reply with quote  #9 
I think the weakness of the "square pie" (which is more like brownies, but I digress) is that it's just some sort of hybrid folded low-rez bar chart. Moreover, this business of filling it in from the corner seems designed to obscure, rather than clarify the data.

When I see some sort of made-up chart type like this, I immediately start looking for what they're trying to hide from me, rather than what they're trying to say.

I suppose one use for the "grid-of-100-squares" layout might be to emphasize that the data consists of integer percentages: 1, 2, …, 99, 100, with no decimal part. And even at that, it's only good for two or at most 3 values in the chart.

But if you're committed to using a square grid of blocks, and if your purpose is to communicate the data and help humans reason about it, then at least fill from an edge.

Which of these images most clearly communicates the value 56% — "a little bit more than half" — (the value shown in the first juiceanalytics example) using the "square of squares" display format?


Here are the kind of gyrations you have to go through to make it possible to quickly scan three values from a square bar chart. How quickly can you compare the three values in these charts (which are, by the way, 58%, 30% and 12%, but you can't always count on having values that can be shown as rectangles)?

Previous Topic | Next Topic
Print
Reply

Quick Navigation: