Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 
mvgadagi
Contributor III
Contributor III

How to handle the very large scale view across left and right Y axis in the Qlik Sense

Hello All,

Can you please help in explaining the better way to put the graph with a view when we have drastic ranges over multiple y-axis measures

Data

Types Sum(Step1_Step2) Sum(Step2_Step3) Sum(Step3_Ste4) Sum(Criteria_1) Sum(Criteria_2)
Type 1 38.1 2.3 2.2 52 3,714.6
Type 2 8.4 0.1 12.8 15 14,495.0
Type 3 1.5 1.7 12.4 69 353.3
Type 4 3.5 0.0 10.0 115 261.9
Type 5 10.4 2.0 2.1 11 5,400.9
Type 6 8.9 0.1 6.1 342 564.0

 

Using this data we wanted to show that Sum(Step1_Step2) Sum(Step2_Step3) Sum(Step3_Ste4) as Stacked bar information and Sum(Criteria_1) as one Y-axis line and Sum(Criteria_2)  as another Y-axis line and problem being the

  1. Sum(Criteria_1) is way smaller compared to Sum(Criteria_2)
  2. Sum of Sum(Step1_Step2) Sum(Step2_Step3) Sum(Step3_Ste4) is way smaller than Sum(Criteria_1) sometimes 

Hence we are not able to fit the Sum(Criteria_1) which is rarely visible proper in the graph and I have attached the graph in the section below, Can you please help me provide a better solution and some kind of hacks where we can solve this problem

 

 

6f052435c95d49649fef71ca082b8411.png

 

Labels (2)
3 Replies
Michael_Tarallo
Employee
Employee

Hi Qlikuser3 - I sent this to a few of my colleagues to see what might be the best way to approach this.

Regards,

Mike T

Qlik

Regards,
Mike Tarallo
Qlik
mvgadagi
Contributor III
Contributor III
Author

@Michael_Tarallo Thanks for looking into this, Please help us with little urgency on this request !!

Patric_Nordstrom
Employee
Employee

Imho, having measures of several orders of magnitude apart is probably not best solved by putting everything in to one chart.

I favor like in this article (https://outlier.ai/data-driven-daily/data-visualization-comparing-multiple-measures-on-the-same-plot...) to reduce to two measures and a dual axis. Another option could be to normalize (if possible) the measures to %.

Thanks,

Patric