Skip to main content
Announcements
See what Drew Clarke has to say about the Qlik Talend Cloud launch! READ THE BLOG
cancel
Showing results for 
Search instead for 
Did you mean: 
Not applicable

cluttered scatter graph

I've put together a scatter graph for 'risks'.  the axis are 'likelihood' and 'consequence' but we only score each on a '1 - 5' scale, so unfortunately, we have many risks with the same score, so although I can show them on the scatter, I wondered if anyone had come across a good way of showing several items at the same place.

On web maps, sometimes they use a technique of showing a symbol with a number in it that shows how many items are at that place, then you hover over it and it expands out to show you the ones underneath.  Maybe something like that?

Sadly, I'm trying to mimic a manually created slide that was used in a meeting and they want it behave the same, so I can't play around with it too much and on the manual one, they would just move the objects around a bit on the PowerPoint slide, so that you could see them all - doesn't seem easy to do that here though.

any ideas welcome!

1 Solution

Accepted Solutions
Not applicable
Author

thanks for that.

But in places I had over 20 items, so even as semi-trnsparent layers, it still didnt really work.

But it did get me thinking and in the end, I added a randum number to each of the two coordinates in the LOAD:

Consequence + (RAND()*0.8) as ConsequenceRAND,

Likelihood + (RAND()*0.8) as LikelihoodRAND

the 0.8 was so that I didnt change the integer value of the point.

Then I used those values in the scatter graph and it seperates them all out really nicely, and becouse they are still in the same 'square' of the graph, it still works.

View solution in original post

2 Replies
swuehl
MVP
MVP

You can play around using the transparent channel (ARGB color code) of the item background color, thus multiple items in the same location should get a darker color than a single item.

Not applicable
Author

thanks for that.

But in places I had over 20 items, so even as semi-trnsparent layers, it still didnt really work.

But it did get me thinking and in the end, I added a randum number to each of the two coordinates in the LOAD:

Consequence + (RAND()*0.8) as ConsequenceRAND,

Likelihood + (RAND()*0.8) as LikelihoodRAND

the 0.8 was so that I didnt change the integer value of the point.

Then I used those values in the scatter graph and it seperates them all out really nicely, and becouse they are still in the same 'square' of the graph, it still works.