3D charts

29 Mar 2006 - 4:03pm
8 years ago
13 replies
886 reads
Robert Hoekman, Jr.
2005

I know I read a nice little rant somewhere recently about how using 3D
charts was unwise when compared to 2D charts that were designed much more
simply, but I can't seem to find any info at all now about why one might be
better than the other.

Personally, I find 3D charts to be way too full of "chart junk" - more
graphics than data - but does anyone here have any more solidifed insights
on the subject? Any info would be much appreciated.

Thanks ...

-r-

Comments

29 Mar 2006 - 4:08pm
Andrew Hedges
2006

For reference, here is the referred-to-rant:

http://presentationzen.blogs.com/presentationzen/
2006/01/2d_or_not_2d_th.html

-Andrew
-----
andrew at clearwired.com / http://clearwired.com/

On Mar 29, 2006, at 3/29/2006 2:03 PM, Robert Hoekman, Jr. wrote:

> [Please voluntarily trim replies to include only relevant quoted
> material.]
>
> I know I read a nice little rant somewhere recently about how using 3D
> charts was unwise when compared to 2D charts that were designed
> much more
> simply, but I can't seem to find any info at all now about why one
> might be
> better than the other.
>
> Personally, I find 3D charts to be way too full of "chart junk" - more
> graphics than data - but does anyone here have any more solidifed
> insights
> on the subject? Any info would be much appreciated.
>
> Thanks ...
>
> -r-

29 Mar 2006 - 4:26pm
Nina Gilmore
2005

Hi Rob,
Distortion of perspective is the number one reason I'd recommend 2D charts over 3D; 3D space makes it too difficult to compare values, volumes and areas to one another. In short, 3D distorts visual relationships and therefore distorts the information. It's also more difficult to simply read the value for a single bar (or line, or area) when depth is factored into the rendering. You wind up having to work harder when translating the graph into meaningful, useful information. Of course, that doesn't prevent people from wanting slick graphs with lots of visual effects on them. But if you look at the graphs commonly shown in news and finance (Wall St. Journal, New York Times, Stock Sites), they are rarely 3D and rarely have visual effects that overpower the data (ok, that's my subjective opinion on a small sample of stuff that I've seen.)

SAP's design guild has a website with graph recommendations (click on "Problems with Charts"):
http://www.sapdesignguild.org/resources/diagram_guidelines/index.html

and Stephen Few's book "Show Me the Numbers" addresses this issue as well (in more detail.)
http://www.amazon.com/gp/product/0970601999/104-0930955-2599130?v=glance&n=283155

Hope this helps.
Best, Nina
-----------------
nina gilmore
interaction design/visual design
business intelligence tools
oracle corporation
-----------------

"Robert Hoekman, Jr." <mmbeta at gmail.com> wrote:
[Please voluntarily trim replies to include only relevant quoted material.]

I know I read a nice little rant somewhere recently about how using 3D
charts was unwise when compared to 2D charts that were designed much more
simply, but I can't seem to find any info at all now about why one might be
better than the other.

Personally, I find 3D charts to be way too full of "chart junk" - more
graphics than data - but does anyone here have any more solidifed insights
on the subject? Any info would be much appreciated.

Thanks ...

-r-

---------------------------------
Blab-away for as little as 1¢/min. Make PC-to-Phone Calls using Yahoo! Messenger with Voice.

29 Mar 2006 - 5:17pm
Jeremy Wood
2006

Interesting question, as I may soon need to make a utility to make
pretty graphs.

Can you give me some examples of what typical "3D" usage is? That may
seem like a silly question, but I want to make sure we're thinking of
the same thing.

For example, I'm picturing "fake 3D", like this:
http://www.jpowered.com/graph_chart_collection/index.htm

... but there's also such a thing as "real 3D" charts, like this:
http://www.idbs.com/xlfit4/index.asp?source=GoogleAdWordsXlfit4
-3DChart-1

In the case of fake 3D, I could make an argument that it's often
overkill. Maybe some subtle depth and shadows are nice -- if
tastefully done -- but you shouldn't add it on principle of adding
something cool.

But if you banned any illusion of depth completely, then you couldn't
even add a drop shadow... and that's generally regarded as a harmless
(if not professional) touch. (As always, no decision is immune to
context, though...).

- Jeremy

On Mar 29, 2006, at 1:03 PM, Robert Hoekman, Jr. wrote:

> [Please voluntarily trim replies to include only relevant quoted
> material.]
>
> I know I read a nice little rant somewhere recently about how using 3D
> charts was unwise when compared to 2D charts that were designed much
> more
> simply, but I can't seem to find any info at all now about why one
> might be
> better than the other.
>
> Personally, I find 3D charts to be way too full of "chart junk" - more
> graphics than data - but does anyone here have any more solidifed
> insights
> on the subject? Any info would be much appreciated.
>
> Thanks ...
>
> -r-
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org
>

29 Mar 2006 - 6:30pm
Nina Gilmore
2005

Hi Jeremy,
I think the question is whether depth is tied to data values (i.e. useful information) or whether the depth is purely for visual effect. The first example you reference is 3D-effect, adding depth for purely visual effect. In Business Applications, I see far more 3D-effect type graphs (using 3D for visual effect) than true 3D graphs. I'm actually not very familiar with true 3D graphs, but know that statistical graphing packages often contain far more varieties of true 3D graphs than business applications do.

Tasteful colors and shading are a good thing! The major limitation I see has to do with display devices and how the graph is going to be consumed. If you're talking handheld devices or black and white printers, less is more. If you're talking projection on an screen (as in presentations), brighter colors and bigger fonts are warranted. I think it all comes down to rendering visual effects without undermining the primary purpose of the graph: communicating information.

Thanks, Nina

Jeremy Wood <jeremy at tech4learning.com> wrote:
Can you give me some examples of what typical "3D" usage is? That may
seem like a silly question, but I want to make sure we're thinking of
the same thing.

For example, I'm picturing "fake 3D", like this:
http://www.jpowered.com/graph_chart_collection/index.htm

... but there's also such a thing as "real 3D" charts, like this:
http://www.idbs.com/xlfit4/index.asp?source=GoogleAdWordsXlfit4
-3DChart-1

In the case of fake 3D, I could make an argument that it's often
overkill. Maybe some subtle depth and shadows are nice -- if
tastefully done -- but you shouldn't add it on principle of adding
something cool.

But if you banned any illusion of depth completely, then you couldn't
even add a drop shadow... and that's generally regarded as a harmless
(if not professional) touch. (As always, no decision is immune to
context, though...).

- Jeremy

---------------------------------
Yahoo! Messenger with Voice. PC-to-Phone calls for ridiculously low rates.

29 Mar 2006 - 7:00pm
Jeff Howard
2004

If you're looking for someone to tear into 3D Charts, you don't have to look much farther
than Tufte's Visual Display of Quantitative Information.

His argument is that one dimensional data shouldn't be conveyed with three dimensions, or
even two, because as humans, we tend to have a hard time seeing through the extra layers.
His examples in the Graphic Integrity chapter revolve around using area or volume to
exagerate differences between simple numbers. He calls it the Lie Factor.

In a pie chart for example, if you skewed it in three dimensions, the slice of the pie closest to
the viewer would look bigger than the slices on the far side of the pie, because of
foreshortening. That prevents the viewer from making accurate comparisons between the
slices... and turns a graphic that isn't much good in the first place into a misleading graphic.

It's worth mentioning that Tufte isn't against 3D charts when there are actually three
dimensions to display. The cover of Visual Explanations is a good example, as are many of
the other spatial graphs he references in his book. Use the dimensions that you need, no
more, no fewer.

Compared to that, adding a drop shadow isn't such a big deal.

// jeff

> Personally, I find 3D charts to be way too full of "chart junk" - more graphics than
> data - but does anyone here have any more solidifed insights on the subject?
> Any info would be much appreciated.

29 Mar 2006 - 8:31pm
Steve Baty
2009

Robert,

Stephen Campbell's 1974 book - "Flaws and Fallacies in Statistical Thinking"
(Dover Publications) has a chapter titled 'Cheating Charts' that discusses
ways in which charts and graphs can (and have) been used to misrepresent and
distort data. He is equally critical of 2D and 3D illustrations, but amongst
his criticisms of 3D charts particularly is the distortion of proportion,
perspective, and scale.

For example, showing a perspective illustration of a house (pictogram) with
a value of $10,000 with, say, dimensions of 1x1x1. Next to that show that
'House prices have doubled' by showing a second house with dimensions of
2x2x2, and a value of $20,000. The result is that the _volume_ of the second
house is actually 8 times the first - not double - leading to the perception
of a much greater increase in the mind of the reader. To be accurate, that
example should have used dimensions of 1.26x1.26x1.26

The same problem can exist in 3D pie charts, where the slices at the 'top'
of the chart are further away from the viewpoint giving the impression of
diminished size. Similarly 3D bar charts where the outlying columns appear
'larger' because you can see all of them rather than having their sides
obscured.

PS: I highly recommend Campbell's book. It contains many, many useful
examples of real-world issues with all forms of statistical analysis, faulty
conclusions based on data, and misrepresentation of data. I keep a copy on
my desk and refer to it regularly. And it's a bargain at $11US.

Best Regards,

Steve Baty
Director, User Experience Strategy
Red Square

On 30/03/06, Robert Hoekman, Jr. <mmbeta at gmail.com> wrote:
>
> [Please voluntarily trim replies to include only relevant quoted
> material.]
>
> I know I read a nice little rant somewhere recently about how using 3D
> charts was unwise when compared to 2D charts that were designed much more
> simply, but I can't seem to find any info at all now about why one might
> be
> better than the other.
>
> Personally, I find 3D charts to be way too full of "chart junk" - more
> graphics than data - but does anyone here have any more solidifed insights
> on the subject? Any info would be much appreciated.
>
>
>

30 Mar 2006 - 1:50am
Jeff Howard
2004

> Does anyone here have any more solidifed insights on the subject?
> Any info would be much appreciated.

Another strategy you could use to argue against 3D charts might be to explore the concept
of conservation of volume. Piaget's experiments showed that young children can't accurately
compare volumes (tall thin glass vs short fat glass). A few years ago the University of Illinois
(UIUC) did a study that suggests we don't get much better at this as we age.

http://www.sciencedaily.com/releases/2003/10/031023070629.htm

Apparently there's also a haptic version of this perceptual shortcoming (that works in sort of
the opposite way) that theaters exploit with the design of popcorn tubs.

// jeff

30 Mar 2006 - 10:53am
Robert Hoekman, Jr.
2005

You know, I've thought about picking up those books about a million times,
but I know that when I do, I'm going to be sucked into a world I'm not ready
to be sucked into. (Who has time to obsess over three books at once?)

Maybe I'll just pick up that one for now, and try to hold off the obsession
for another time. Thanks for the summarization, though. Very helpful.

-r-

On 3/29/06, Jeff Howard <id at howardesign.com> wrote:
>
> If you're looking for someone to tear into 3D Charts, you don't have to
> look much farther
> than Tufte's Visual Display of Quantitative Information.
>
> His argument is that one dimensional data shouldn't be conveyed with three
> dimensions, or
> even two, because as humans, we tend to have a hard time seeing through
> the extra layers.
> His examples in the Graphic Integrity chapter revolve around using area or
> volume to
> exagerate differences between simple numbers. He calls it the Lie Factor.
>
> In a pie chart for example, if you skewed it in three dimensions, the
> slice of the pie closest to
> the viewer would look bigger than the slices on the far side of the pie,
> because of
> foreshortening. That prevents the viewer from making accurate comparisons
> between the
> slices... and turns a graphic that isn't much good in the first place into
> a misleading graphic.
>
> It's worth mentioning that Tufte isn't against 3D charts when there are
> actually three
> dimensions to display. The cover of Visual Explanations is a good example,
> as are many of
> the other spatial graphs he references in his book. Use the dimensions
> that you need, no
> more, no fewer.
>
> Compared to that, adding a drop shadow isn't such a big deal.
>
> // jeff
>
>
> > Personally, I find 3D charts to be way too full of "chart junk" - more
> graphics than
> > data - but does anyone here have any more solidifed insights on the
> subject?
> > Any info would be much appreciated.
>
>
>

30 Mar 2006 - 11:01am
Robert Hoekman, Jr.
2005

Indeed, there is quite a difference. I'm talking more about the 3D charts
that only use depth for visual effect. See the attached for a specific
example.

I'm trying to justfiy getting this thing redesigned. Any ammo I can get is
great. Thanks to everyone for the great information. You guys are always so
helpful!

-r-

On 3/29/06, Jeremy Wood <jeremy at tech4learning.com> wrote:
>
> [Please voluntarily trim replies to include only relevant quoted
> material.]
>
> Interesting question, as I may soon need to make a utility to make
> pretty graphs.
>
> Can you give me some examples of what typical "3D" usage is? That may
> seem like a silly question, but I want to make sure we're thinking of
> the same thing.
>
> For example, I'm picturing "fake 3D", like this:
> http://www.jpowered.com/graph_chart_collection/index.htm
>
> ... but there's also such a thing as "real 3D" charts, like this:
> http://www.idbs.com/xlfit4/index.asp?source=GoogleAdWordsXlfit4
> -3DChart-1
>
> In the case of fake 3D, I could make an argument that it's often
> overkill. Maybe some subtle depth and shadows are nice -- if
> tastefully done -- but you shouldn't add it on principle of adding
> something cool.
>
> But if you banned any illusion of depth completely, then you couldn't
> even add a drop shadow... and that's generally regarded as a harmless
> (if not professional) touch. (As always, no decision is immune to
> context, though...).
>
> - Jeremy
>
> On Mar 29, 2006, at 1:03 PM, Robert Hoekman, Jr. wrote:
>
> > [Please voluntarily trim replies to include only relevant quoted
> > material.]
> >
> > I know I read a nice little rant somewhere recently about how using 3D
> > charts was unwise when compared to 2D charts that were designed much
> > more
> > simply, but I can't seem to find any info at all now about why one
> > might be
> > better than the other.
> >
> > Personally, I find 3D charts to be way too full of "chart junk" - more
> > graphics than data - but does anyone here have any more solidifed
> > insights
> > on the subject? Any info would be much appreciated.
> >
> > Thanks ...
> >
> > -r-
> > ________________________________________________________________
> > Welcome to the Interaction Design Association (IxDA)!
> > To post to this list ....... discuss at ixda.org
> > List Guidelines ............ http://listguide.ixda.org/
> > List Help .................. http://listhelp.ixda.org/
> > (Un)Subscription Options ... http://subscription-options.ixda.org/
> > Announcements List ......... http://subscribe-announce.ixda.org/
> > Questions .................. lists at ixda.org
> > Home ....................... http://ixda.org/
> > Resource Library ........... http://resources.ixda.org
> >
>
>
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: chart.jpg
Type: image/jpeg
Size: 69265 bytes
Desc: not available
Url : /pipermail/discuss-interactiondesigners.com/attachments/20060330/5c1e7a42/chart.jpg

30 Mar 2006 - 3:31pm
John Vaughan - ...
2004

RH > I'm trying to justfiy getting this thing redesigned. Any ammo I can get
is great.

RE: Your Example 3-D Chart

As a general rule, the admonishments from other participants that "Less is
More" are very well-taken, esp. in the case of this example chart. There
are a bunch of problems here - A couple of them are 3-D-related.

Superfluous use of color distracts from the overall pattern that you're
trying to communicate.

This chart shows Sales Dollars over the course of 8 months. The bold color
difference among the month bars adds no information - and actually distracts
us with unnecessary "info-noise". Use just a single color (or a consistent
color theme).

Less vibrant color is also helpful. A slight differentiation in shade could
be used to indicate that "this value is above/below [a threshhold]" - IF
that info is useful.

Are the numeric values important? If so, they are being overshadowed by the
designy-ness of the 3-D bars. You may want to reduce the visual impact of
the bars and/or emphasize the presentation of the numeric values in order to
get your point across.

The depth/thickness of the bars is a bit much (and it chews up a lot of
valuable horizontal space). The size and darkness of shadowed "extra" area
in the 3-D shape visually distorts the proportions of the bar values. Less
bulkiness - or even just a simple 2-D dropshdow - is safer.

The bar titles at the bottom (the "month-year") are textual overkill that
distract from the info. Thetextisallsquishedtogetherandhardtoread. A
simple month heading - with a separate indication of the year - does the
job.

V
e
r
t
i
c
a
l

t
e
x
t

s
u
c
k
s

Especially if you've got a bunch of words. ESPECIALLY if it's the title of
the chart. Do it normal. Your audience will figure it out.

All this is IMHO. Hope this helps.

30 Mar 2006 - 3:45pm
Robert Hoekman, Jr.
2005

I

A
g
r
e
e

c
o
m
p
l
e
t
l
y.

Thanks for the great response!

-r-

On 3/30/06, John Vaughan <vaughan1 at optonline.net> wrote:
>
> RH > I'm trying to justfiy getting this thing redesigned. Any ammo I can
> get
> is great.
>
> RE: Your Example 3-D Chart
>
> As a general rule, the admonishments from other participants that "Less is
> More" are very well-taken, esp. in the case of this example chart. There
> are a bunch of problems here - A couple of them are 3-D-related.
>
> Superfluous use of color distracts from the overall pattern that you're
> trying to communicate.
>
> This chart shows Sales Dollars over the course of 8 months. The bold
> color
> difference among the month bars adds no information - and actually
> distracts
> us with unnecessary "info-noise". Use just a single color (or a
> consistent
> color theme).
>
> Less vibrant color is also helpful. A slight differentiation in shade
> could
> be used to indicate that "this value is above/below [a threshhold]" - IF
> that info is useful.
>
> Are the numeric values important? If so, they are being overshadowed by
> the
> designy-ness of the 3-D bars. You may want to reduce the visual impact of
> the bars and/or emphasize the presentation of the numeric values in order
> to
> get your point across.
>
> The depth/thickness of the bars is a bit much (and it chews up a lot of
> valuable horizontal space). The size and darkness of shadowed "extra"
> area
> in the 3-D shape visually distorts the proportions of the bar
> values. Less
> bulkiness - or even just a simple 2-D dropshdow - is safer.
>
> The bar titles at the bottom (the "month-year") are textual overkill that
> distract from the info. Thetextisallsquishedtogetherandhardtoread. A
> simple month heading - with a separate indication of the year - does the
> job.
>
> V
> e
> r
> t
> i
> c
> a
> l
>
> t
> e
> x
> t
>
> s
> u
> c
> k
> s
>
> Especially if you've got a bunch of words. ESPECIALLY if it's the title
> of
> the chart. Do it normal. Your audience will figure it out.
>
> All this is IMHO. Hope this helps.
>
>
>

30 Mar 2006 - 4:21pm
cfmdesigns
2004

>From: Jeremy Wood <jeremy at tech4learning.com>
>
>Interesting question, as I may soon need to make a utility to make
>pretty graphs.
>
>Can you give me some examples of what typical "3D" usage is? That may
>seem like a silly question, but I want to make sure we're thinking of
>the same thing.

The "bad" stuff is where the 3-D at best doesn't add value and at worst, hides info or misleads the people reading te chart.

Unquestionably, if you are dealing with 3 variables of info rather than 2, a 3-D chart or some sort is probably called for. All 3-D charts aren't bad.

-- Jim Drew
Seattle, WA

30 Mar 2006 - 5:37pm
Josh Galban
2005

Here's one more mark against 3D charts: they often don't hold up as well as
2D charts when printed, faxed or photocopied in black & white.

Charts intended for printing -- whether 3D or 2D -- must work well in black,
white and a limited range of grays. The extra chartjunk found in a 3D chart
gets compressed into that more limited palette. It's easier to design a 2D
chart for everyday printing, faxing or photocopying by just leaving that other
mess out.

--Josh

------ Original Message ------
Received: Thu, 30 Mar 2006 08:02:33 AM PST
From: "Robert Hoekman, Jr." <mmbeta at gmail.com>

[ snip ]

I'm trying to justfiy getting this thing redesigned. Any ammo I can get is
great. Thanks to everyone for the great information. You guys are always so
helpful!

-r-

Syndicate content Get the feed