Column Charts and Bar Charts are more or the less the same thing; both represent data using rectangular columns / bars. The major difference between them; Column Charts present the data on vertical columns and Bar Charts on horizontal bars.
This answer to this depends on two things.
- the data
- the available work-space
However the most important thing to consider and decide is which presents the data in the most useful and meaningful way.
It would be great if we could define some hard set rules around which chart to use, but the fact is it depends! However there are a handful of things we can consider that should help us to decide.
Just a quick note at this point; the charts shown above have been created at a 16:9 ratio, which reflects the screen resolutions we generally use today. Clearly creating a taller narrower chart changes things significantly, but we generally work with shorter wider charts.
The length of our category axis labels
As can be seen in the above examples, Bar Charts better accommodate longer category axis labels (in this case country names) without compromising the usability of the chart. Longer category axis labels on Column Charts start to eat into the valuable and limited vertical space which makes the values represented by the columns less discernible.
As in the above Column Chart example; longer category axis labels can be shown at an angle with the ideal been 45 degrees in many cases. Applying more angle than this means we start to use up the vertical chart space (this also results in the category axis labels becoming difficult to read; and in sore necks for that matter); Applying less angle and we start to make the chart difficult to read particularly when trying to associate values to their labels.
I would therefore recommend using a Bar Chart to represent data with long category axis labels.
The number of category axis labels
This ones a little difficult as some people appear to think differently to me. I believe that a Column Chart is better for displaying more category axis labels; however others believe its the Bar Chart that’s better at this.
In the above examples at least we have more horizontal space than vertical space which suggests the Column Chart can accommodate more category axis labels than the Bar Chart (in this case 10 country names); However the length of the category axis labels also comes back into play at this point.
I would therefore recommend using a Column Chart to represent data with a higher number of category axis labels; unless your dealing with long category axis labels in which case a Bar Chart may be better – trial and error at this point!
The size of the values
Its not uncommon to have charts representing data in the millions, billions and even trillions (not my bank balance though unfortunately) so we need to think about how to best represent such large values.
In the above examples we can see that the Column Chart better accommodates larger axis value labels on the y-axis than the Bar Chart does on the x-axis. This could be overcome by showing the axis value labels on an angle on the Bar Chart AND/OR, as we’ve done in the above example, showing the axis value labels in a larger unit (Millions in this case). However for the same reason mentioned previously, we don’t want to take up too much of that vertical space by showing the axis value labels too vertically on a Bar Chart.
So a Column Chart is better for displaying large values then? well not necessarily! The bars on a Bar Chart generally have more room to play than the columns on a Column Chart (assuming a ratio of 16:9 or similar is used). The values represented by bars on a Bar Chart are generally easier to decipher than the columns on a Column Chart; the difference between the bars (and therefore the difference between the values) can also be more easily deciphered on a Bar Chart.
So this again is a little difficult to conclude, but I would recommend using a Bar Chart to represent data with larger values (and using larger unit and/or angled value axis labels if required).
This one can probably be discussed and debated all day long; so as I mentioned at that start go with what you think represents your data best (not what looks prettiest but what presents the data in the most useful and meaningful way).