topic Multiple groups (clusters) in spiderchart? in QlikView

Multiple groups (clusters) in spiderchart?

Sun, 18 May 2014 13:54:36 GMT

I have a data model that distinguishes organizations by size, e.g. number of employees. In a spiderchart, I want to present the results clusters of the number of employees of organizations: organizations with fewer than 1.000 employees, between 1.000 and 2.000 and more than 2.000 employees.

In a spiderchart, I want to present the average results for each cluster. How do I do this? when I'm not using the clusters, i.e. the total of all organizations, the figure is. When I want to present the results for each cluster, no data is displayed.

Expression used:

if ([Total employees] <= 1000, count ({<[Function supported] = {"Yes"}>} [Function supported]))

count (TOTAL <[Specification function - Importance for System]> [Function supported])

See screenshot of figure: now displayed with the total. What I would like to have is a line for each cluster of employees.

Re: Multiple groups (clusters) in spiderchart?

hic — Sun, 18 May 2014 17:45:47 GMT

Add a second dimension (Calculated) with

=If([Total employees]<=1000, 'Group A',

If([Total employees]<=2000, 'Group B',

'Group C'))

HIC

Re: Multiple groups (clusters) in spiderchart?

Sun, 18 May 2014 19:14:05 GMT

I added a calculated dimension as described above.

=If([Total employees]<=1000, 'Group A',

If([Total employees]<=2000, 'Group B',

'Group C'))

The expression I used is thus:

count ({<[Function supported] = {"Yes"}>} [Function supported]))

count (TOTAL <[Specification function - Importance for System]> [Function supported])

Unfortunately, the object only shows results for group C. However, there are also outcomes in group A. They are strangely not shown.

Important thing to mention: I want to show the percentage of 'yes' on a 100% scale, in relation to the number of organizations that has answered 'yes' or 'no' (by function). 'Yes / no' should be excluded.

Re: Multiple groups (clusters) in spiderchart?

hic — Sun, 18 May 2014 20:26:03 GMT

OK, then you need to make your calculated dimension a little more complex. You probably have a field "Organization" or something similar, and an Organization can only have one possible number of [Total employees]. If so, try

=Aggr( If(Only([Total employees])<=1000, 'Group A',

If([Total employees]<=2000, 'Group B',

'Group C')),

Organization)

HIC

Re: Multiple groups (clusters) in spiderchart?

Mon, 19 May 2014 20:40:49 GMT

Your assumptions are correct. I have a field called 'organization'. I've worked out your idea in the properties of the object ('dimensions'). Still only cluster C is shown, while there are also data (number of employees) in cluster A and B. I have adjusted the ranges (for testing), where:

= Aggr (If (Only ([Total employees]) <= 100000, 'Group A'

If ([Total employees] <= 200000, 'Group B'

"Group C")),

Organization)

Still only results of cluster C is shown (or this might be the outcome of all organizations?). I'm going to try it some more, based on your suggestions. Thanks to far.

Re: Multiple groups (clusters) in spiderchart?

hic — Mon, 19 May 2014 20:53:47 GMT

To debug, put the expression in a list box to see what you get. To test further, change the expression to use Avg() instead of Only():

=Aggr( If(Avg([Total employees])<=1000, 'Group A',

If(Avg([Total employees])<=2000, 'Group B',

'Group C')),

Organization)

Alternatively, you can try the Class function:

=Class( Aggr( Avg([Total employees]), Organization), 1000 )

HIC

Re: Multiple groups (clusters) in spiderchart?

Tue, 20 May 2014 19:09:19 GMT

Yes! Now I get the results on the screen as I would expect! Thank you Henric Cronström. It worked with the use 'Avg' in de expression.