Control Chart

[Kushal_Chawda]

What is Control Chart?

The control chart is a statistical analysis used to study how a process changes over time. Data are plotted in time order. Control charts are used to routinely monitor quality or simply you can say, control chart is process improvement technique.

What is a process?

A process is simply what we do. It can be filling out an expense report, checking a person into a hospital, burning a CD with music, driving to work, filling a prescription, etc. All these process generates the data. We take these generated data and plot on control chart to improve efficiency.

There are many types of control chart, but we are going to use the basic control chart with Std dev method to understand the concept.

Followings are some scenario where we can use the control chart

1) In manufacturing to test if machinery are producing products within specifications.

2) suppose you want to reduce the time it takes to admit someone to the hospital.

3) If you want to know how views are going up or down for particular website.

Components of Control Chat

1) Avg line

2) UCL (Upper Control line) = Avg Value + Std Deviation* Std Dev Multiplier

3) LCL (Lower control line) = Avg Value - Std Deviation* Std Dev Multiplier

So, basically we can see which are the points "Out of Control" or "In Control".

You can control your Std Dev multiplier using variable.

How to create the Control Chart.

1) Create the Combo Chart.

2) Take any Time dimension of your interest which is suited for your analysis, here I have taken the MonthYear. Also    make it Continuous in Axis setting

3) Create two Expression, one for Line and second for Symbol.

4) Create expressions for Avg line, UCL & LCL which will be used as reference lines in the chart. Check the Chart properties ->Presentation for reference line setting.

5) Create the Background color setting for Symbol expression

if( sum(Sales) < vUCL and sum(Sales) > vLCL, LightGreen(), LightRed())

6) Finally the chart will look like below

There are some other useful links available to create the control charts

QlikView Control (SPC) Chart

See also :

Benford's law

Bollinger Bands

[danielgargiulo]

We have released a Control Chart extension for Qlik Sense.  More information and a 60 day trial can be downloaded from here.  

Key features include: • Outliers Highlighted • Runs ‘above’ and ‘below’ the average highlighted • Trends up and down highlighted • Moving average based on Runs and/ or Trend triggers

Key configuration options include: • Customise Labels for Dimensions, Measures and Rules • Customise Line and point style for Dimensions, Measures and Rules • Pick the number of points to set the highlighting criteria for both Runs and Trends

For more information email: extensions@acumenbi.co.nz .

[peterderrington]

Could i just clarify something......

The value for vStdDev is set here as "2"

Is this a standard value or does it differ dependant on other factors?

Thanks.

[Kushal_Chawda]

This is standard deviation and it differs based on requirement. By definition, standard deviation 2 covers the 95% spread of your data. You can use 3 as well to get more spread of your data.

[peterderrington]

Thanks for coming back to me, much appreciated. Just wasn't sure what the 2 or 3 equalled when determining standard deviation.

Thank you.

[qliky88]

Hey @Kushal_Chawda great post thanks for your help 🙂 

From your own experience do you think this can work with the OOTB combo chart for Qlik Sense? 

I want to use the OOTB one as opposed to the Vizlib one due to the Y-Axis automatically adjusting for values, however I cannot seem to locate an option to add a reference line in the OOTB version. 

Thanks in advance!

[Mhidz]

Thank you Mr. Kushal. 

Please help on how to do this in Qlik Sense?

Thanks,

Mhidz