Quartile Logic!!

[amar_never_dies]

Hi All,

Can anyone please share the quartile logic in qlikview.

Load * INLINE [

Sales,Year

500, 2013

10,2014

300,2015

400, 2013

1000,2014

30000,2015

50, 2013

1050,2014

700,2015 ];

Total rows=8

quartile is fetching the particular Sales row value when sorted in ascending order.

25th Quartile=8*0.25=2nd value i.e 2nd row sales value in ascending order.

Output Sales(25th Q)=50

Output Sales(50th Q)=8*0.5=4th value=400

Output Sales(75th Q)=8*0.75=6th value=700

I tried using fractile function but put is not as expected.

can anyone provide the logic. Please note fieldvalue works only on load order. i want sales to be sorted in ascending order first, any function or logic will be a great help!!!

thanks

[petter]

The total number of rows that you have is 9 ... not 8. So maybe you get things to add up if you revise accordingly.

[amar_never_dies]

petter thanks for replying. 8 was by mistake. consider 8 rows and share the logic accordingly for implementing quartile as explained above.

[petter]

I simply calculated it directly in the GUI with the help of an InputBox with different expressions. This could of course be done both in the load script but also in a particular chart. The expressions will be the same.

[petter]

If you are happy with the answer - please close the thread by marking it as answered ...

[amar_never_dies]

hi Petter,

I think i have already shared the output. the desired output should be as before.

fractile function does not returns the expected result.

moreover, i want a bar chart in which i want quartile values relative to year like 2012,2013,2014,2015.

for e.g

2015 year- sorted quartile values at 0.25,0.5,0.75,1 and so on for 2013, 2014...

[petter]

I can't get the reference data right so have a look at this Excel-file where I calculate Quartiles in different ways. I can't reproduce the numbers you have - this is what I get (included the Excel-file for reference):

So which numbers do you consider right for what you want in QlikView. QlikView Fractile()-function works as the .INC (inclusive) functions in Excel....

[amar_never_dies]

hi Petter,

Load * INLINE [

Sales,Year_Test

500, 2013

10,2014

300,2015

400, 2013

1000,2014

30000,2015

50, 2013

1050,2014

700,2015

1500, 2013

15465,2014

6700,2015

50545, 2013

105,2014

70,2015

883,2015

];

Sorry for late replying. Consider the above data set and i want the output in qlikview 3 values at 25 quartile, 50 quartile and 75th quartile which are the corresponding 4th, 8th and 12th row values after sorting . I want it to be dynamic. I have great thousands rows of data year wise. Means if i select year 2013 then it shows corresponding 25%,50% and 75% quartile values in year 2013. For Full data set, i am sharing the expected output, But needs it to be dynamic on selection as well.

Total rows in Sales- 16

16*0.25=4th row value of sales after sorted ascending=105

16*0.50=8th row value after sorted ascending=700

16*0.75=12th row value after sorted ascending=  1500

Expected Output: Quartiles in table box/straight table

Year            Sales

2014             105  

2015             700

2013            1500

Please remember that it has to be dynamic as well on selection.

Suppose if year 2015 is selected then the output should be 2nd !, 3rd and 5th values of year 2015 data set. like:

Year           Sales

2015           300

2015           700

2015           6700

Please help any way you can!!!!! thanks

[petter]

There are different methods of calculating quartiles. Just look at Wikipedia about quartiles and it will explain the differences quite clearly. The differences are very clear with small data sets - so what you prescribe as the correct on you tiny data set is correct in one method and different in other methods. The important thing is that the differences disappear when a certain amount of numbers is used.

Using the Fractile() function in QlikView is quite easy - just remember that the results you get are based on a particular method and comparing it with another tools function and the results might give different answers - and especially when you have few numbers to calculate over.