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  
jhcarrell

Registered:
Posts: 17
Reply with quote  #1 

My understanding ….

 

In bar graphs the one axis serves to separate data points, while the other provides the measuring stick for those data points (i.e. bars).  Additionally, the data area of the chart is defined (framed) by the two axi.

[image]

However, I’ve recently seen where some people have started electing to forgo the measure axis in the name of simplicity by instead opting to direct label each bar.  As shown below:


[image]

While direct labeling does give us a more precise reading of the exact value of a given bar, do you feel anything is lost by doing this?


__________________
"In the realm of data visualization, we face many challenges that are worth pursuing. Creating an effective radial gauge is not one of them. It is a fool’s errand. Do I strike you as a fool?" - Stephen Few
danz

Registered:
Posts: 183
Reply with quote  #2 
Is not a totaly wrong method or so.

I personally consider that value labels written directly on chart area would make me focus on reading values rather than visually comparing bars (purpose of the bar chart).

Also, when the scale is properly designed, besides the regular value decoding purpose, it helps me roughly estimate the distribution of values based on scale ticks/bins. Just reading the values would tell me I have four values between 10 and 20 and two values above 20. Using the scale I see that suggested bin is 5%, I have three values between 10 and 15, one value between 15 and 20, two values between 20 and 25 but no values between 0 and 5 or between 5 and 10. I have no idea how relevant is this for that graphic, just saying that a scale can show more insights.

When I want to display all values I prefer to have a table like design with three columns where first column is the category label, second is the value formatted with the required precision and third is the bar chart.
jlbriggs

Registered:
Posts: 191
Reply with quote  #3 
@danz I completely agree.

I find the axis scale is an important reference, and I think much context is lost by removing it.

If you need to show the specific numbers, doing it outside of the bar chart, accompanying the category labels, is my preferred method (whether or not in an actual tabular set up).

I also find that with no scale, and labels on the bars, I have to read each label to understand the scale - with the axis properly labeled, I see the full range, sequentially and evenly ordered, and can much more quickly and fully grasp the range of data.
sfew

Moderator
Registered:
Posts: 805
Reply with quote  #4 
Similar to danz and jlbriggs, I believe that the axis with a quantitative scale provides a useful framework for rapidly understanding the values, which works better than labeling the values for each bar. I'm not aware of any actual research that has been done to confirm this, but my intuition, honed by many years of experience, suggests that this is true. I have on rare occasions labeled the values and eliminated the axis and scale with small sets of values, but I avoid this as a general practice.
__________________
Stephen Few
infinite8s

Registered:
Posts: 4
Reply with quote  #5 
Hi Stephen,

If possible, could you go into the reasons you eliminated the axis and scale on those rare occasions?
sfew

Moderator
Registered:
Posts: 805
Reply with quote  #6 
infinite8s,

I was afraid that someone might ask me this question. In truth, I regret including a handful of examples of bar graphs with labeled values and without a quantitative scale along the axis, because there probably is no good justification for every omitting the scale. At best I can say that, with a small set of values, labeling the bars with values and omitting the scale does relatively little harm and can be less intimidating for people who are not comfortable reading graphs.

__________________
Stephen Few
Previous Topic | Next Topic
Print
Reply

Quick Navigation: