1

I am looking for a smart way to visualize proportions, or ratio, over time in a chart. Since the relation between the variables is more interesting than their individual value, I can't just plot the variables individual values on a line chart, at least not in any way I have tried yet.

One of the problems I am facing is that none of the variables represented in the chart should give the impression of being more "preferable" that the other variables.

All our other visualizations over time are represented with line charts but I can't seem to get this right when it comes to proportions/ratios.


I have two cases where I am looking for a solution for this problem:

Case 1: Scenarios where there are two variables, A and B.

  • A + B will always result in 1.

Case 2: Scenarios where there are more that two variables, i.e. A, B, C and D.

  • A + B + C + D will always result in 1.

Here is an image with some example data: enter image description here

I divided them into two different cases since I figure there might be different approaches for more variables than two.

4
  • What do you mean by "Smart way"? Can you give some examples of the data you are going to display? Just referring to them as variables is not that helpful.
    – SteveD
    Jul 5, 2016 at 11:36
  • @Splatz I updated the post just now. I hope this gives anyone trying to answer this a better understanding of what I am trying to achieve. Thank you for your comment!
    – user86540
    Jul 5, 2016 at 12:01
  • Thanks. What is the goal of the visualisation? Is it as simple as just to see how the ratios have changed over time?
    – SteveD
    Jul 5, 2016 at 12:07
  • @Splatz The goal is exactly as you guessed: to show how the ratios have changed over time.
    – user86540
    Jul 5, 2016 at 12:30

3 Answers 3

3

I think this is the only visualisation that meets your need for Case 1. Here you can clearly see a declining pattern:

enter image description here

Your case 2 is more difficult to solve using this pattern because now you are looking to compare 4 things and the stacked bar approach does come with some challenges when it comes to comparing a trend over time. Consider this example - you can see it is much more difficult to spot individual variable trends compared to how easy it is with just 2 variables:

enter image description here

If it is an important goal to compare the relative differences between variables for each year, then perhaps this visualisation approach will be better (it is so much easier to see which variable is best - assuming this is your goal):

enter image description here

Note I made this up very quickly but you would need to ensure the 4 bars if stacked would equal 1.

2
  • If it is important to see the values for A and B you could superimpose the values in the bars
    – SteveD
    Jul 5, 2016 at 12:33
  • Thank you very much for your earlier comments and the awesome answer. I totally agree with the stacked bar chart when using many variables, it gets a little hard to spot the relative differences. We are starting with the visualization with two variables and you answer helped us come to a decision. Thanks again!
    – user86540
    Jul 5, 2016 at 13:50
3

Sounds like you want a stacked percentage chart. This can be either a bar chart, or a line-based one.

enter image description here

enter image description here

1
  • 1
    Thanks alot for your answer! After some internal discussions we decided to go with the stacked bar chart.
    – user86540
    Jul 5, 2016 at 13:46
-1

Stacked line chart is the way to go on this. PixelSnader, in their answer above, ends with a great example of a stacked line chart. The stacked nature helps visually distinguish between the series (A,B,C, and D). And the line chart format is easier to asses changes over time, which is why a bar chart might be a close second in this situation, but often a bit chaotic.

2
  • 1
    Please provide more detail in your answers. Explain why your suggestion is the correct course of action, and if possible provide an example.
    – JonW
    Apr 11 at 14:20
  • Sorry JonW, brand new here. Apr 12 at 21:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.