New to QlikView

Discussion board where members can get started with QlikView.

Announcements
QlikWorld, June 24-25, 2020. Free virtual event for DI and DA gurus. Register Now
cancel
Showing results for
Did you mean:
Highlighted
Contributor III

How to identify the values ​​at the bell curve extremes?

HI!

How can I show how are the employees (field cod_empl) out of the normal distribution of the curve?

Is it  possible thay if no year is selected show the curves of all years?

Are the expressions ok or there is something wrong?

8 Replies
Highlighted
MVP

Re: How to identify the values ​​at the bell curve extremes?

>>How can I show how are the employees (field cod_empl) out of the normal distribution of the curve?

Define the extremes. Do you mean more than 2 standard deviations from the mean? Or do you mean by fractiles?

>>Is it  possible thay if no year is selected show the curves of all years?

That's that way Qlik behaves out of the box.

Logic will get you from a to b. Imagination will take you everywhere. - A Einstein
Highlighted
Contributor III

Re: How to identify the values ​​at the bell curve extremes?

Hi!

In my first question, i mean standar deviations from te mean, so when a value from the list box assig is selected, a table box or a straight table show who are outside the normal distribution and their values.

In my second question  I was referring to that if no year is selected, a bell curve for each year is shown at the same time. In my chart example, if no year is selected, only one bell curve is show

My thrid question,  I mean that if, for example, t2 is selected, the bell curve has a strange shape since it does not reach the axis of the dimensions completely.

Highlighted
MVP

Re: How to identify the values ​​at the bell curve extremes?

It's still not clear what you mean by "outside the normal distribution". A normal distribution runs from -infinity to +infinity, so technically nothing can be outside. Practically your data will only occupy a subset of that range. It is common to consider data within a certain number standard deviations of the mean, which will in turn set what fractile of the data is included (if the data is truely normally distributed). The remaining items outside that std deviation range are then considered outliers. Do you mean something along those lines?
Logic will get you from a to b. Imagination will take you everywhere. - A Einstein
Highlighted
MVP

Re: How to identify the values ​​at the bell curve extremes?

>>In my second question  I was referring to that if no year is selected, a bell curve for each year is shown at the same time. In my chart example, if no year is selected, only one bell curve is show

Then add Year as a dimension to the chart

Logic will get you from a to b. Imagination will take you everywhere. - A Einstein
Highlighted
MVP

Re: How to identify the values ​​at the bell curve extremes?

>>My thrid question,  I mean that if, for example, t2 is selected, the bell curve has a strange shape since it does not reach the axis of the dimensions completely.

Real data may not be normally distributed, and if you have a small data set, you may not have enough data points to get a bell shape.

A normal distribution is a common pattern for the general behaviour of attributes in a population, but any real world sample can only approximate the normal distribution, or it may not be normal at all. For example waiting times in a queue are not usually normally distributed.

Logic will get you from a to b. Imagination will take you everywhere. - A Einstein
Highlighted
Contributor III

Re: How to identify the values ​​at the bell curve extremes?

Thanks for sharing your knowledge, really didactic!!

Ok, based in yoru explanation, i need to show the persons in  the " remaining items outside that std deviation range are then considered outliers".(positive and negative)

Abusing a little of your kindness, when you refer to fractiles, do you mean the value that I define in the list box Desv?

Really apreciate your interest in helping me.

Highlighted
Contributor III

Re: How to identify the values ​​at the bell curve extremes?

I must be doing something wrong.
Adding year as dimension in my chart, it only shows a single curve.
Maybe It will not be possible to do it according to my data model
Highlighted
Contributor III

Re: How to identify the values ​​at the bell curve extremes?

Ok! I understand. Thanks for the explanation. Your example is very clear