SPC Chart Analysis

Problem

Process dimensions, temperatures, and defect counts often reduce to averages and ranges, making it hard to tell whether a process is in statistical control, whether variation is drifting, or where out-of-control points originate. This skill focuses on manufacturing process data and turns an Excel-to-report SPC workflow into a single script call covering common control charts, rule-based exception detection, and capability calculations.

How It Works

It first identifies whether the data is variable or attribute type, then selects among Xbar-R, Xbar-S, I-MR, P, C, and U charts. It then runs scripts/spc_analysis.py, with core capabilities including:
- Chart generation: produces an HTML report with SVG charts.
- Nelson rules: detects anomalous points and explains the triggering rules.
- Stability and capability assessment: computes Cp/Cpk when USL and LSL are supplied.
- Historical comparison: compares recent mean, standard deviation, anomaly counts, and stability changes.

Boundaries

It is well suited to repeated monitoring of one characteristic, such as part diameter, temperature, or batch defect rate. It is not intended for mixing different characteristics in the same conversation; start a new session when the data object changes. Without specification limits, the analysis focuses mainly on process stability rather than conformance to engineering tolerances.