Skip to content

User Guide

Task-oriented guidance for using mfgQC day to day. If you want the formula and the source standard behind a method, see the Reference instead.

  • Install — pip install mfgqc, supported Python versions.
  • Quickstart — the load → spec → analysis flow on a worked dataset.
  • Reading the assumption report — what each guardrail checks, what a warning means, and when to opt into auto-correction.
  • Choosing a control chart — the inference rule and how to override it.
  • Gage R&R workflow — roles, variance components, %study vs %tolerance, ndc, and the AIAG verdict.
  • The audit workflow — record, export, and verify a result's lineage. This is the page that shows the differentiator end to end.

The one idiom

Everything in mfgQC follows the same shape: a verb produces an object, and the object has methods.

import pandas as pd, mfgqc

qc  = mfgqc.load(df, measure="width", subgroup="lot", subgroup_size=5)
qc  = qc.spec(lower=1.0, upper=2.0, target=1.5)   # attach metadata fluently
cap = qc.capability()                              # run an analysis

cap.report()              # full text: numbers + assumption checks + recommendations
cap.summary()             # a flat dict of the headline scalars
cap.to_dict()             # full JSON-serializable payload (consume this from code)
cap.view(save="cap.png")  # the canonical chart

Once you know this shape, every analysis in the library works the same way.