Solved! Go to Solution.
This is an old thread, but I needed to create some XmR control charts recently, and thought I'd leave this here in case someone needs it in the future. The attached workflow creates X (individual) and mR (moving range) control charts.
Note, this is only flagging points outside of the control limits (it's not checking any other WECO or Nelson rules).
An alternative option would be to use the R tool with one of the SPC packages (such as qcc or rspc) to apply the rules, and either chart in R or output the flagged data from the R tool to be charted in Alteryx.
Thank you. This is very helpful.
Does anyone have an alteryx workflow for SPC?
Can anyone QA my SPC Alteryx workflow?
coo, cant believe how short that workflow is and does all that!
attached is my attempt at a run chart rules and is MASSIVE in comparison
this applies a test to a run chart, testing for shifts, trends and the correct amount of runs given the amount of data points. I've tried to code the tests from the article here. cant code one of them, the "astonomical data point" as Alteryx seems to lack an Eyeball tool.
.
The workflow outputs a text summary of the checks performed. The last 5 yards arent complete in terms of what youd want output, but the tests seem to be accurate.
I want to add a thing to "fix" the median if it is ran and passes all 3 tests, so the median isnt re-calculated every time a new datapoint is added, causing potential "drift"
Any errors spotted will be gratefully received.
i am looking for users that have an SPC macro that can be used for any input files and that automatically output XMR charts.
Hi Kelly,
I love this workflow and the control charts that were developed from it. What was the use case that your workflow answered?
Thanks so much!
@chriswilliams41 - I don't remember the application for XmR charts specifically, but the general use case for control charts was monitoring data volume and data quality. My role is in data curation, so I was using a suite of control charts to monitor the size and content of my data loads (with emailed alerts if parts of the process went out of control). We've since adopted an enterprise data quality tool, so I've been able to retire my homegrown monitoring system!
(off topic: I greatly enjoyed your presentation on the AMP engine at the recent Portland user group!)