which logic is used for bubbles' colour in bubble chart?
Hello,
I try to understand how the colours of the single bubbles are determined in a bubble chart. What is the logic behind the colours ?
I make an example. I have numeric values, actual and plan for sold quantity in units. I build the numeric algorithm progress = actual - plan.
I set the bubble chart as below, and 1 entity by row (the customer).
I set the same block (progress in units) for both Y axis and for the colour .
I mean, the colour should have a similar behaviour than the Y axis. I have the feeling, the colours are random...
The BOARD help doesn't explain the logic of the colours, except in the case where there are two entities by row, which is not my case. The only element is:
optionally a fourth value can be used to color the bubbles.
Does anyone know how the colours of the bubbles are determined by BOARD when a specific block is set for it ?
Answers
-
Hi interested typ,
although I'm not quite sure about this, my assumption would be that BOARD uses the value you choose for colouring to determine the value of the colour it uses for displaying the associated bubble in a way that the lowest value will get the lowest colour-code.
Since the colour-codes are restricted to a certain amount (depending on your CID) this could mean that once BOARD has walked through all available codes it will start again with the lowest level if the maximum number of colours has been reached to colour the next higher value.
Hope this helps.
Kind regards,
Helmut
0 -
Hello,
what does CID mean ?
I did several tests and I think your assumption is correct. Thank you very much !
I created a new dummy cube and set it for both parameters Y axis and bubbles colour of the bubble chart.
- I fed the cube with random values for each customer, but only integers from 1 to 10. Result is :
- I fed the cube with random values, but only integers from 1 to 50. Result is:
So when a KPI has continuous (or many discretes) values , which is almost always the case, the bubbles' colours make no sense on a visualization perspective, the colour doesn't bring any useful information.
I will - instead of setting a block for the colour - set a second entity by row of the layout, to keep the colour meaningful.
0 - I fed the cube with random values for each customer, but only integers from 1 to 10. Result is :
-
Hi interested typ,
CID means "Corporate Identity Design(er)".
And you're right--either a second entity on the row or an entity to colour the bubble (remember to use the entity's code as the value). The next thing is, that the colours don't stay the original way after you've made a selection to one or two entities; that's why I opened the idea Keep dimensional colours in Portfolio Object .
Kind Regards,
Helmut
1 -
once BOARD has walked through all available codes it will start again with the lowest level if the maximum number of colours has been reached to colour the next higher value.
Ideally, I would have expected that BOARD combines two information before deciding which bubble receive which colour:
- number of colours available, let's say m= 300. These colours must be sorted by code from 000000 to FFFFFF.
- number of distinct values (n) of the block used as colouring parameter
two possibilities:
- m >= n. it is easy, BOARD split the codes available to each value.
- m<n, BOARD has to process a quantization of the n values, so that at the end the number of quantiles are <= m. Then all bubbles should receive the colour code of the quantile they belong to.
in the case of BOARD, I am no sure whether the colours are sorted , according to my screenshot with 50 discrete values, I can see it is unfortunately not the case. This example shows as well that BOARD doesn't use all codes extensively from 000000 to FFFFFF and doesn't sort them. So even in case a), the colour information is not very useful. In case b), I would have expected several bubbles having exactly the same colour because of their belonging to a given data quantile.
it's a pity.
0