Statistical Quality Analysis Tools with Plain-Language Interpretation

Statistical tools built for quality engineers

A Cpk in a spreadsheet gives you a number, not whether the customer will accept it. Ten statistical tools built for manufacturing quality: enter your data, get the result with capability and Gauge R&R graded against the usual acceptance thresholds, and ask for a plain-language reading of what it means and what to do next.

Start a 30-day trialSee example documents

Capabilities

What Analyze does

Capability Studies (Cpk/Ppk)

Calculate short-term Cpk and long-term Ppk with bias-corrected sigma, graded acceptable at 1.67 and up, conditional from 1.33, marginal from 1.00, and unacceptable below that.

Measurement System Analysis

ANOVA-based Gauge R&R with repeatability, reproducibility, and number of distinct categories.

Statistical Process ControlCore

X-bar and R, X-bar and S, I-MR, p, np, c, and u charts with Western Electric rules. Detect shifts, trends, and out-of-control conditions.

ANOVACore

One-way, two-way, and multi-factor analysis of variance, with Tukey and Bonferroni post-hoc tests, to identify significant sources of variation.

Design of ExperimentsCore

Full and fractional factorial designs with main effects and interaction analysis.

Hypothesis Testing & RegressionCore

T-tests, z-tests, proportion tests, chi-square, and simple, multiple, and polynomial regression with confidence intervals and p-values.

Pareto Analysis

Identify vital few vs trivial many defect categories with the 80/20 rule.

Acceptance Sampling

AQL sampling plans per ANSI/ASQ Z1.4 (single sampling, normal inspection, general level II) with OC curves and lot evaluation.

Six Sigma Metrics

DPMO, sigma level, yield, and process performance calculations.

Who does what

Statistical analysis: the software prepares, an engineer decides.

Prepared by the software

  • Explains a Cpk/Ppk result in plain language with recommendations; the low, medium, or high risk level comes from fixed thresholds, not from the model
  • Explains Gauge R&R findings in plain language with actionable improvement suggestions
  • Flags Western Electric rule violations on control charts with fixed rules, and explains their practical significance on request
  • Generates full and fractional factorial run sheets from the factors and levels you enter

AI interpretation is available on Core plan and above.

Decided by an engineer

  • Approves or rejects each document change the software proposes from a corrective action or an engineering change, before it is applied
  • Approves, rejects, or waives each PPAP element, and approves or rejects the submission
  • Approves NCR dispositions and signs off each corrective action
  • Signs control plans, PFMEAs, and other title-block documents by role, and enters the authorized signature on the warrant

Signed in the app: typed signature, name, date

Workflow

How it works

  1. 01Select analysis type
  2. 02Enter or upload data
  3. 03Run calculations
  4. 04Review AI interpretation
  5. 05Export results

Use cases

Built for data-driven quality engineering

  • Quality engineers running initial process capability studies for PPAP submissions
  • Metrology teams conducting Gauge R&R studies on measurement equipment
  • SPC coordinators monitoring process stability with control charts
  • Process engineers designing experiments to optimize manufacturing parameters
  • Six Sigma practitioners calculating DPMO, sigma levels, and process yields
  • Quality analysts performing hypothesis tests and regression analysis on quality data

Questions

Frequently asked questions

What is process capability (Cpk) and why does it matter?
Process capability index (Cpk) measures how well a manufacturing process stays within its specification limits, measured to the nearer limit, so an off-center process scores lower than a centered one. Customers commonly ask for 1.67 on a new process or first study and 1.33 on an established one. QualityEngineer.ai calculates short-term Cpk and long-term Ppk using bias-corrected sigma and grades the lower of the two: 1.67 and up is acceptable, 1.33 to 1.67 conditional, 1.00 to 1.33 marginal, and below 1.00 unacceptable. On Core and above, an AI model can then explain the result in plain language.
What is Gauge R&R and when do I need it?
Gauge R&R (Gauge Repeatability and Reproducibility) measures whether your measurement system is capable of detecting the variation it needs to measure. It separates variation into repeatability (same operator, same part) and reproducibility (different operators). You need a Gauge R&R study for any measurement system used in capability studies or SPC; it is required by IATF 16949 and industry MSA methodology. The platform performs ANOVA-based Gauge R&R and reports %GRR, number of distinct categories, and adequacy assessment.
What SPC charts does the platform support?
Seven chart types: X-bar and R, X-bar and S, individuals and moving range (I-MR), and the p, np, c, and u attribute charts. Western Electric rules 1 through 4 are on by default: a point beyond the 3-sigma control limits, nine points in a row on one side of the center line, six points in a row rising or falling, and fourteen points in a row alternating up and down. You can switch on all eight rules. The rule checks are fixed calculations; on request, an AI model explains what the violations suggest, such as a process shift, a trend, or a mixture.
How does AI interpret statistical results?
After a calculation, ask for an interpretation and an AI model returns a plain-language explanation of what the numbers mean, specific recommendations for improvement if needed, and a low, medium, or high risk level. The risk level is set by fixed thresholds, so the model cannot talk a failing result into a pass. This turns statistical output into quality decisions, even for engineers who are not statistics experts. Interpretation is available on Core and above.
What other statistical tools are available?
Beyond Cpk/Ppk, Gauge R&R, and SPC, the platform includes: ANOVA (one-way, two-way, and multi-factor, with post-hoc tests), DOE (full and fractional factorial designs with interaction analysis), hypothesis testing (t-tests, z-tests, proportion tests, and chi-square), regression analysis (simple, multiple, and polynomial regression with p-values), Pareto analysis (80/20 rule for defect categorization), acceptance sampling (AQL plans per ANSI/ASQ Z1.4), and Six Sigma metrics (DPMO, sigma level, yield calculations).

Related

Works with

Try it on a real part.

30-day trial, no card required. Bring your own drawings and documents, or start with the sample part we set up for you.

Start a 30-day trialTalk to us