topic Re: discard outliers in the scatter chart in QlikView

discard outliers in the scatter chart

gauthamchilled — Wed, 25 Nov 2020 16:16:04 GMT

hi,

i would like to discard or ignore the outliers in the scatter chart. can anyone help me? attached the example

Re: discard outliers in the scatter chart

gauthamchilled — Fri, 03 Mar 2017 04:00:11 GMT

any help?

Re: discard outliers in the scatter chart

muthukumar77 — Fri, 03 Mar 2017 04:37:45 GMT

Hi,

Do you want like this?

Re: discard outliers in the scatter chart

gauthamchilled — Fri, 03 Mar 2017 05:17:55 GMT

I dont see any difference between my file and your image.

I want to exclude the values which are unusually large on both axis..

Re: discard outliers in the scatter chart

tresB — Fri, 03 Mar 2017 05:33:36 GMT

May be like in attached sample.

Re: discard outliers in the scatter chart

gauthamchilled — Fri, 03 Mar 2017 05:47:51 GMT

Hi Tresesco,

This is a sample, so we could easily figure out what we dont want and filter it out.

But actual dataset with many countries and many years of revenue and profit, we cant hardcode the number,

Is there anyway to remove the outliers based on the current selected value dynamically using function like fractile or standard deviation???

Re: discard outliers in the scatter chart

gauthamchilled — Fri, 03 Mar 2017 06:31:42 GMT

bump

Re: discard outliers in the scatter chart

tresB — Fri, 03 Mar 2017 06:43:59 GMT

You have to have a rule to decide which values to be discarded, say for example if you wish to discard the top two values, you may try expression like:

Sum ({<[Customer Name]={"=Rank(Sum (Sales))>2"}>}Sales)

Update: Using fractile()

Sum ({<[Customer Name]={"=Sum(Sales)<Fractile(total Sales, 0.99)"}>}Sales)

Re: discard outliers in the scatter chart

jonathandienst — Fri, 03 Mar 2017 07:03:15 GMT

Often you will find outliers by using the Fractile() function. For example, include only the best 90%, use something like:

=If(Sum(Value) > Fractile(Sum(Value), 0.05) And Sum(Value) < Fractile(Sum(Value), 0.95), Sum(Value))

Re: discard outliers in the scatter chart

abhishek_singla — Fri, 03 Mar 2017 07:53:13 GMT

Hi Gautham Prasad,

What do you define outliers?

I've filter out the outliers based on following formula:

As per Inter quartile range (IQR), outliers are those whose value is

1. value > Q3 + 1.5*IQR

2. value < Q1 - 1.5*IQR

See below pic.

I've done this in qlik sense. PFA .qvf file.