Data for XmR / SPC / control charts

[charlie_taylor]

XmR / SPC / control charts are a popular format to track data and detect abnormal variations - spikes and trends. Most online examples refer either to dedicated software or Excel templates (e.g. http://www.spcforexcel.com/knowledge/control-chart-basics/purpose-control-charts).

This is something Tableau should be able to do - and it's possible but it's very manual and not flexible enough (http://www.tableau.com/public/blog/2013/11/control-charts-2236).

I'm looking for examples of automated data preparation for production of XmR charts so that it can be done quickly for a large number of KPIs. Is this something Alteryx could do / has done?

 

[AndyC]

Hi @charlie_taylor

It looks like this post was missed originally.  I've asked the Solutions Engineering team if anyone has used Alteryx in this way before.

Thanks

[brendafos]

I am interested in Alteryx and Control Charts.... a reply would be good.

[ecastruita]

Also Interested in this topic as it is a must-have capability for any worthwhile organization's optimization processes.

[markstephens]

Another vote for this

But also, for run chart rules, see attached doc

Basically something that could check for

1. shifts
2. trends
3. max/min number of runs.

[napplewh]

I need to do SPC on large datasets is this possible using an Alteryx workflow 

[markstephens]

Why yes it is but youll have to build it yourself.

i did one for run charts, to detect shifts (6 or more  above or below median) and trends (5 or more all on one direction), min/max number of runs

Its the same logic to be able to do control chart rules.

just using formula to compare previous datapoint with current and testing the run chart rules in a column that adds one for every time theres a consecutive direction in the run. 

Bit hacky but it works.

[Husker59]

Mark,

Can you share the Run Chart Alteryx file?

Also, how do you perform a WINDOW_AVG in Alteryx?

Thanks you,

Ross

[napplewh]

is it possible to share your workflow as i also need to do a more advanced version of SPC to detect shifts and trends

i've created a manual one myself that calculates Sigma, median and outliers above or below the Upper and Lower Control limits.

I am happy to share my version to the community if anyone is interested or could QA my SPC workflow.

[Husker59]

Hello,

I would love to see your version.

Thanks in advance.

Ross LeGrande

rlegran@its.jnj.com