Anything with a tolerance can be judged against it: a fill weight, a machining dimension, a settlement time, a price band, a margin floor. The histogram in ATX AnalyticsTable takes a lower and an upper specification limit and a target, shades what falls outside, and adds Cp, Cpk and the share out of specification in percent and parts per million. They are computed from the same rows the chart draws, on your own computer or database, and they go into the PDF report with the chart. This page also says plainly what it is not: a statistical process control (SPC) suite with control charts and subgroups.
Add the widget with Dashboard → Add Widget → Histogram, or select a column of numbers and click Reports → Charts → Histogram on a sheet. In the properties panel, Specification limits and statistics takes LSL, USL and a target.
Cp = (USL − LSL) / 6σ and Cpk = min(USL − mean, mean − LSL) / 3σ, where σ is the sample standard deviation of every value the chart counts. There are no subgroups and no within-subgroup estimate of σ, so in the terms of the AIAG manual and of Minitab these figures are overall, long-term capability: what those tools label Pp and Ppk.
That is the right reading for a period of production data, a month of settlements or a portfolio of prices. Read it as how the process actually performed over the rows shown, not as what it could do if special causes were removed.
Cp and Cpk from the overall standard deviation, with out-of-spec ppm, on any histogram with spec limits.
In the app, up to 100,000 values are read and binned exactly. Counting in the database returns only counts per fine sub-bin and exact sums, so a table of any size costs the same, no row-level data leaves the source, and the mean and standard deviation, and with them Cp and Cpk, stay exact. The SQL is written for DuckDB, PostgreSQL, SQL Server, Oracle and BigQuery.
If you need control charts (X̄-R, X̄-S, individuals, p or c charts), run rules such as Western Electric or Nelson, within-subgroup Cp and Cpk, gauge R&R or non-normal capability models, use a dedicated SPC package next to it. ATX AnalyticsTable is for putting capability into the same governed reports, schedules and dashboards as the rest of the business's figures.
Cp compares the width of the specification with the spread of the process and ignores where the process is centred. Cpk also takes the centre into account: it measures the distance from the mean to the nearer limit, so an off-centre process scores lower.
They are computed from the overall sample standard deviation of all values shown, with no subgroups, which is what the AIAG manual and Minitab call Pp and Ppk. The chart labels them Cp and Cpk, as most business users expect.
Yes. Any measure with a tolerance works: a settlement or delivery time, a price band, a margin floor, a response time. The out-of-spec share in ppm reads the same way.
No. There are no control charts, run rules or subgroup-based indices. It brings capability figures into scheduled reports and dashboards next to the rest of the business data.
This site needs JavaScript for the full experience. Every page above is reachable without it.