4 Replies Latest reply: Feb 20, 2012 11:54 AM by josljosl

# Multiple cummulative distribution functions (CDFs) in one chart

Currently, I can create a cumulative distribution by using

Dimension:

Class(my_value,1)

my_product

Expression:

count(my_product)

tick Relative & Full accumulation.

This works fine for one CDF so probability goes from  0 to 100% but if I add another my_product, it will split the count between those two,

so I will have 1 CDF going from 0 - 60% second one from 0 - 40% which adds up to 100%...but I want them both to go to 100%.

Cheers

Message was edited by: josljosl Adding example charts, the good one goes from 0 - 100%, the bad one is only up to 0 - 60%.

• ###### Multiple cummulative distribution functions (CDFs) in one chart

Hi Jos

It's the second dimension my_product that is causing the two lines. If you get rid of it, you will not have this problem and the cumulative amount will be the total for both products.

Erica

• ###### Multiple cummulative distribution functions (CDFs) in one chart

Thanks Erica,

if I do remove the [my_product] from the dimension, it will result in showing only 1 CDF if I select both apples and bananas.

What I try to see is a CDF for each one of them going up to 100%.

Do you know of any other way to achieve this ?

Cheers

Jo

• ###### Multiple cummulative distribution functions (CDFs) in one chart

Sorry Jos, I wasn't clear what you wanted.

The "relative" button only calculates the distribution relative to the entire total of the expression. This is equivalent to

[unticking relative]

=count(my_product)/count(total my_product)

In this case your chart is calculating

[cumulative apples] / [total cumulative apples and bananas] as a %

resulting in the unusual total %.

To fix this, you need to do the expression manually, then divide by the total of just the fruit. To do this, use the <> brackets after the total to total by just the type of fruit.

Untick the relative box, and enter in

=count(my_product)/count ( total <my_product> my_product)

as the expression. This should solve your problem.

Regards,

Erica

• ###### Multiple cummulative distribution functions (CDFs) in one chart

Thanks Erica, that did the trick !