ds@mcpanalytics.aiA data scientist you can email your data and question to, and get a reliable response. Not AI slop.A data scientist you can email. Not AI slop.

Is your process in control?
Drop in your data and see the control chart

Put your measurements in one file, in the order they were taken. See the control chart with limits drawn from your own data, the points that break the standard rules, and, with specification limits, how capable the process is.

Try it on example data

Which chart this draws

An individuals chart: one measurement per row, such as a fill volume per batch or a daily reading. The centre line is the process mean. The limits sit three sigma either side, with sigma estimated from the average change between consecutive readings, divided by 1.128. A moving range chart goes with it.

It does not draw X-bar and R, p, np, c or u charts.

Three questions this answers

  1. Is the process stable, or is something beyond ordinary variation at work?
  2. Which readings broke a rule, and which rule?
  3. With specification limits, can the process hold them?

What your file needs

  • One row per measurement, with the measured value in a numeric column.
  • An order column, such as a date, a time or a sequence number. Without one the tool uses the order the rows arrive in and says so.
  • At least 20 measurements. Below that the tool stops, because it needs them to estimate its own limits.
  • Optional: a lower and an upper specification limit, each as a column holding the same value on every row.

What you get back

  • The chart with its centre line and control limits, and the moving range.
  • The points flagged by each of four rules: one beyond the limits, nine in a row on one side of the mean, six rising or falling in a row, and two of three beyond two sigma.
  • With specification limits: Cp and Cpk from the short-term spread, Pp and Ppk from the overall spread, and the share of readings outside specification.
  • A normality check, because the capability indices assume a normal distribution and mislead when it fails.
  • The method named and the R code.

Questions people ask

Which control chart should I use?

Use an individuals chart when each reading stands on its own, as with one measurement per batch or per day. That is the chart this tool draws. Grouped measurements and counts of defects need other charts.

What are the out-of-control rules?

The tool runs four: a point beyond the three sigma limits, nine in a row on the same side of the mean, six in a row steadily rising or falling, and two of three in a row beyond two sigma on the same side.

What is the difference between Cp and Cpk?

Cp compares the width of the specification with the spread of the process. Cpk also takes into account how well the process is centred between the limits, so it is never higher than Cp.

How many points do I need?

At least 20. The limits are estimated from the data itself, and the tool stops with a message below that.

What does out of control mean?

That a signal appeared that ordinary variation alone would rarely produce. It tells you to look for a cause. It does not say the product is out of specification.

Have your measurements ready?

Try the example file first, then drop in your own.

Try it on example data

Related