Comments
A2 Academy: Get Visual, Add Value
View Comments: Newest First | Oldest First | Threaded View
Re: Data visualization from Hell
  • 10/25/2013 7:07:19 PM
NO RATINGS

@Lyndon_Henry    Thanks for this example of a "Pie Chart Gone Wild" .  How is one to make any sort of sense much less a decision based on this ?   And the sad thing is - this kind of garbage is out there.

Re: "Analytics needs Storytelling."
  • 10/25/2013 7:02:59 PM
NO RATINGS

@CandidoNick    I agree and I can certainly relate.  Nothing better than a well crafted story to get difficult concepts accross.  Analytics does need stories, but first it has to find some storytellers.

Data visualization from Hell
  • 10/25/2013 6:22:40 PM
NO RATINGS

..

Sorry about the rude title... but nothing else seemed to capture the idea.

Anyway, here's a pie graph example that I think is closer to what Jonathan was warning against:

 

There are so many slices to this that it's hard to make any real sense of the proportions. You can't readily compare even similar slices to one another in terms of magnitude -- even the two biggest pieces (Retail Trade and Government). They look fairly close, but if you examine the actual values, Government is a lot bigger. So, what does this do for us except confuse, baffle, or induce retinal glazing?

Incidentally, this is Red State Montana. What would the public think if they realized that their single biggest sector of employment was ... Socialism (i.e., Government)?

 

Re: Change of plans
  • 10/25/2013 5:51:21 PM
NO RATINGS

..

Beth writes


 Jon certainly isn't the only one to bash the 3-D pie. I've heard that time and again -- and I do see why. While you've been very deliberate in the use you show here -- you understand the data and how you'd like the business users to view the results -- that doesn't hold true for everybody else.


 

Well, I get the point.

Anyway, this discussion inspired me to go looking for really bad examples of data visualization -- i.e., the quest for "the graph from Hell...". I know I've seen a bunch, but actually the search was harder than I thought.

But I did come across this one:

 

 

Anything that makes me stare at it with open mouth for more than a minute, trying to figure out what it's trying to convey, fails the "What is the reason for data visualization?" test, and falls in my search category.

Incidentally, this is supposed to be graphing (i.e., "visualizing") the following:

Composition of number of business enterprises researchers by field of research (2005)

 

Re: Change of plans
  • 10/25/2013 10:28:34 AM
NO RATINGS

Hi Lyndon, Jon certainly isn't the only one to bash the 3-D pie. I've heard that time and again -- and I do see why. While you've been very deliberate in the use you show here -- you understand the data and how you'd like the business users to view the results -- that doesn't hold true for everybody else. How often do you suppose the 3-D pie chart is merely the result of somebody playing around in Excel rather than truly thinking about the end result and how viewers will perceive the data? I'm guessing a lot. Analise Polsky raised this during a session the other day at The Big Data Conference, and gave a good example. You can see that here: It's a Big Data Jungle Out There

Re: Change of plans
  • 10/24/2013 10:16:26 PM
NO RATINGS

..

Beth writes 


Hyoun Park has had to bow out of today's lecture. But still do join in, as A2 blogger Jonathan Schwabish, a US policy analyst and data visualization creator, is going to jump in and deliver today's class instead. Teaching about data visualizatin is a passion of his, so this will no doubt be a great session.


 

It was indeed a very interesting giscussion. One of the things we discussed was problems in presenting data in graphs, and Jonathan seemed a bit disdainful of pie graphs.

Personally, I do find pie graphs useful. Also, I try to use colors strategically. I also think a 3D effect can lend a bit of pizzazz if it's not overused.

Anyway, here's an example of a pie graph I used in a presentation. 

Notice how the magenta really dominates? That's on purpose...

 

 

Change of plans
  • 10/16/2013 8:37:01 AM
NO RATINGS

Hyoun Park has had to bow out of today's lecture. But still do join in, as A2 blogger Jonathan Schwabish, a US policy analyst and data visualization creator, is going to jump in and deliver today's class instead. Teaching about data visualizatin is a passion of his, so this will no doubt be a great session.

 

Re: "Analytics needs Storytelling."
  • 10/15/2013 7:14:58 PM
NO RATINGS

I would add one thing that is key to successful visualization. It has to resonate for it to have full impact.

"Analytics needs Storytelling."
  • 10/15/2013 6:19:15 PM
NO RATINGS

'Storytelling' is how the best professors from my college days got their points across. Integral to passing info, but it has to be done right.



INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +