3

What kind of data visualization can show 3 aspects of the data including time (the others are the number of conversions and revenue)? For example, I thought about bubble chart where size is one aspect and Y position is another, another example is a multi-set bar chart

enter image description here enter image description here

3

It depends on the tradeoffs (and the story you need to convey): You can break the charts into adjacent, separate displays, or change the chart to emphasize relative change over time.

Break into separate displays:

This tradeoff means you have precision about each figure, but need to parse 2 data points. You can try to amend this with interaction with either chart:

enter image description here

Try making it like a sentence:

enter image description here

In this case, since the y-axis' are separate, hovering on one chart gives you interactivity in the next. I've seen this done vertically as well.

It might be harder to parse the trendline, but the tradeoff is accuracy. Bubble charts can be difficult, as area is hard to parse.

Another solution: if change over time (not volume) is important, change the y-axis to reflect that:

enter image description here

What about dual y-axis charts?

There used to be more of an emphasis on dual y-axis visualization, but they've fallen out of favor due to perception problems:

Why not to use two axes, and what to use instead

The proportions of the two scales are often different from each other in dual axis charts. If the left axis would go down to zero, the chart would be twice as long. If the right axis would go down to zero, the chart would be almost three times as long.

| improve this answer | |
  • Thanks, @Mike, but I don't want to compare the number of purchases and revenue over time, I want to see them both as different aspects of the same data representation, like in the bubble chart, I see size as revenue and Y position as number of purchases – Eran Bar May 27 at 7:44
  • @EranBar do you want to see the correlation of (ideally) a smaller number of conversions producing the highest revenue? I would think that's a good starting point that x number of customers are bringing in much more than during a previous period. What are users doing next with this info? can they run detailed reports to investigate? – Mike M May 28 at 0:47
  • Yes, either high or low revenue per conversion on a timeline, and you can drill down to the data (so yes, its hi-level), at the moment, I tend to think that bubble chart is the best option. – Eran Bar May 31 at 19:51
1

Without seeing the data or what correlations you want to show it's difficult to say. Sometimes you just have to see what visualization method suits the data and the users by "trying them on".

It sounds like you might benefit from trying the data in violin plots or stream graphs?

There's a good selection to review here: http://visualizationuniverse.com/charts

| improve this answer | |
  • Thanks, amigo, what I am asking is what visualizations to test on, I need one, like the bubble chart, that can show 2 non-related aspects of the data: number of purchases made and revenue, on top of a timeline. – Eran Bar May 27 at 7:40
0

I'm making some hypothesis, but I suppose revenues are positively correlated to time (the longer, the more money). The conversion rate is a side information for most of decision-makers as long as the money made is big.

The bubbles also have some sense of "area" but time*something else is a strange unit, that doesn't make much sense.

Revenues/time areas

The top rectangle is short, high revenues/time and high conversion/rate

The bottom rectangle is long, smaller revenues/time and medium conversion rate

| improve this answer | |
  • No, these are 3 separate dimensions – Eran Bar Jul 13 at 6:25

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.