1

I'm a bit stuck here. I need to represent about 40 variables in a bar chart. Each variable needs to be visible... it would be fine if each ended on a different part of the x axis, but unfortunately ~36 of the variables end in the same location. Does anyone have any tips on how to display it more easily/clearly, so that it's not a giant list of bars at the same length?

Edit: Thank you all for the responses! They did help. The photo was just a chart generated by a program to show a better idea of the amount of information. The colors and labels it has are irrelevant. I cannot give too many details as it is confidential.

  • When looking at your numbers I'm wondering what is the bar chart supposed to show? What's the point oft the chart? Bar charts visually display numbers relative to each other. Your numbers are all the same so there's nothing to display. A bar chart doesn't work on numbers like these because there's no difference in them. It turns out to be a list. – moot Nov 22 '19 at 17:22
  • What do they represent?? If they came out all the same maybe recheck what the source data actually means and if a bar chart carries any utility in displaying your data, perhaps another technique? – mastablasta Nov 27 '19 at 5:33
0

I would pick up on Michael's point about questioning whether you do need to show all those variables

1) The hexadecimal labels are pretty much unreadable by humans anyways - they're not much use

2) Why do the labels alternate, are they actually pairs of values (in which case show differences?) or is there just not space for every label?

3) Ask users why they're looking at this data - if they're only looking for data out the ordinary, why not just show the outliers and compact everything else to 1 category/row?

4) Or taking that further, compact all the rows down to just showing unique values, the labels then show the value and item count, and have tooltips displaying all the hex values that share that value?

5) Taking it even further, show all the values as one boxplot with whiskers and outliers - you lose the labels but as said they're not human readable. Label the outliers if need be.

4

The only thing that I can think of is to break up the 40 variables you have into groups, and apply the small multiple visual design concept to create small groups of visual representation.

There are many examples where this data visualization design concept can be used effectively to break up the monotony of a large block of information that needs to be displayed, while still conveying the same overall information in more digestible chunks.

Here is an example of how raw data can be structured for display in different visual formats, including the small multiple.

enter image description here

Of course, without really knowing your design constraints or the intended audience, the best thing I can suggest is to question whether you need to present all of the information on the screen at the same time, and if it is actually easier for the audience to absorb and understand the information you want to present.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.