Is this Set Analysis?

[Anonymous]

In an earlier post (now deleted) I made a poor job of explaining my needs.  Now I've heard of Set Analysis, and perhaps I can clarify.

I have a table containing the number of people in a region, grouped by Ethnic group, Age5 (age, grouped in 5-year bands), and Sex for each TLA (local authority) of interest.  I want to be able to use this figure as the denominator in an incidence study.

I want to restrict Qlik Sense's modifications of the denominator to these four groups.  Any other modifier will be the numerator.

So: my data looks like this sample:

TLA	Ethnic group	Age5	Gender	Pop
001	European	00-04	Female	441
001	European	00-04	Male	513
001	European	05-09	Female	1161
001	European	05-09	Male	1173
001	European	10-14	Female	291
001	European	10-14	Male	288
001	European	15-19	Female	48
001	European	15-19	Male	39

If I select a TLA in my dataset, the denominator should sum Pop for that TLA.  I need to be able also to select an ethnic group, Age5, or Gender, and use the total for this selection as the denominator.  So: these elements are allowed to change, but this is the only modification allowed.

Elsewhere in my dataset I will modify the numerator by selecting an accident type and location.

I should be able to count the number of car crashes, for example; I should be able to modify these by the ethnic group, age group or gender of the person involved.

What I'm looking for is this: how do I show the number of car crashes involving European males aged 15-19 in a TLA, divided by the population size for European males in the TLA?  In a normal application the Pop figure is modified by the number of European males in the TLA.

I believe I'm looking for a Set Analysis, but I don't understand how my Set Analysis can group the denominator according to my four variables.  I'd appreciate any help!

[OmarBenSalem]

You can have your fields as filters, just drag and drop them in your sheet:

Then, if you want to only have a KPI object with the information based on your selection; it would have as an expression:

count(carCrashes) / sum({<AccidentType=,location=>}Pop)

With this, you'll have a count of car crashes that depends on your selections in all of the field divided by the population that depends on your selection but that won't be affected by your selection of the AccidentType and location.

[vinieme12]

Good reads for reference

https://community.qlik.com/docs/DOC-4951

https://www.analyticsvidhya.com/blog/2014/01/set-analysis-qlikview/

[daveamz]

Hi Steve,

Yes, the set analysis should do the work:

What I'm looking for is this: how do I show the number of car crashes involving European males aged 15-19 in a TLA, divided by the population size for European males in the TLA?  In a normal application the Pop figure is modified by the number of European males in the TLA.  

Something like:

In a table with one dimension ('TLA') use the following expression:

Sum({<[Ethnic group] = {'European'}, Age = {'15-19'}, Gender = {'Male'} >} Pop) /

Sum({<[Ethnic group] = {'European'}, Gender = {'Male'} >} Pop)

This should give you an idea of how you can use set analysis. Be aware that set analysis can be used in conjunction with filters (you can ignore certain filters, force others etc.)

Set analysis is very powerful and it is worth investigating the documentation. You can find a lot of resources here, on community in the blog post section.

Regards,

David

[Anonymous]

Thank you Omar, Vineeth and David!

I believe I now have on my screen pretty much what I'm looking for, and you were all most helpful.

Our injury data allows us to filter by many fields, and for the moment I'll limit these to just three while I continue to learn.  At the moment, this dataset is looking very good.

Thank you,

Steve