Live control chart
Drop in a CSV. The chart, the control limits and the out-of-control points are worked out in your browser.
Try it
The sample is loaded. Swap in your own.
Special-cause signal detected. 1 point breached the control limits: the process isn't stable, so investigate before judging capability.
Run this on a real project, with the charter alongside
This demo runs in your browser and your data never leaves this page. In SigmaForge, DataForge saves the analysis to your project and DocuForge drafts the charter from it.
How to read it
Two questions, in the right order.
What an I-MR chart tells you
Whether the process changed. The individuals chart plots each reading against limits set by its own variation; the moving range underneath it is what those limits are built from. Points inside the limits are noise. A point outside is a signal worth chasing.
What Cpk tells you
Whether a stable process fits inside the spec. Cpk compares the distance from the mean to the nearer spec limit with three sigma. Below 1 the process is making parts outside the spec; 1.33 or more is the usual mark of a capable one.
- Read 01
The center line
The average of your readings, in the order you took them.
CL = mean
- Read 02
The control limits
Where the process’s own variation says a reading should fall. They are worked out from your data, never typed in.
mean ± 2.66 × average moving range
- Read 03
A point outside
A signal that something changed at that moment. Find out what happened before you judge capability.
special cause
- Read 04
Cpk
Give it spec limits and it shows how far the mean sits from the nearer one, in units of three sigma.
1.33 or more: capable
Inside DataForge
Every result, saved with the project.
In the product the chart is one of 32 analyses. It runs on data pasted, uploaded, read from a screenshot or pulled from the project, the figures come from the engine, and the result is saved as a deliverable of the phase it belongs to.
- 28 of 32 analyses computed by the engine
- 22 chart types, from I-MR to CUSUM
- A written reading of every result, and what to do next
The chart family
When the individuals chart is not the right one.
- Control · Green Belt
I-MR control chart
Individuals and moving range, with out-of-control signals. The chart on this page.
- Control · Green Belt
X-bar and R chart
For readings taken in subgroups, with the subgroup means and ranges on two charts.
- Control · Green Belt
Attribute charts
p, np, c and u charts, for counts of defects and defectives.
- Control · Black Belt
EWMA chart
Weights recent points more heavily, to catch a small shift that persists.
- Control · Black Belt
CUSUM chart
Adds up deviations from target, so a slow drift shows sooner.
- Measure · Yellow Belt
Process capability
Cp and Cpk against your specification limits, with a non-normal version for skewed data.
More in resources
- All resourcesThe free tools, and what each one answers
- Cpk calculatorFree, runs in your browser, no account
- White paperProof, not promises: AI in improvement work you can audit
- Case studiesThree worked examples, start to finish, on demo data
- VideosThe product on video, from 45 seconds to the full story
- Verify a certificateCheck any certificate by its ID
- MethodologiesLean, DMAIC, A3, Kaizen, root cause analysis
Your own data
Bring the chart that keeps you up at night.
Thirty minutes, on a spreadsheet of yours: the chart, the capability, the document and the gate.