Documentation / SPC / Control chart selection
I — Free decision instrument
Which control chart should I use?
Describe the measurement and sampling design. SVEND recommends the chart, explains the boundary, and can run a documented example through the same Meganeura engine used by the Analysis Workbench.
Free · No account · No saved data · Seven Shewhart chart families
Recommendation
I-MR chartOne continuous observation arrives at a time, so there is no rational subgroup whose within-group variation can be estimated.
Keep the observations in time order. Control limits describe process behavior; specification limits describe product requirements.
II — Run a teaching dataset
Each example is deterministic synthetic manufacturing data, released CC0. Its page metadata names the intended lesson and the tempting wrong chart.
III — Run a small dataset
Privacy boundary: The submitted table is sent to SVEND for this calculation, limited to 2,000 rows and eight numeric columns, and is not saved. Raw values are excluded from application logs and analytics. An example dataset is available where submitting data is not acceptable.
IV — Engine result
V — Selection reference
| Chart | Use when |
|---|---|
| I-MR | Continuous measurements arrive individually. |
| X-bar/R | Continuous measurements form small, fixed rational subgroups. |
| X-bar/S | Continuous measurements form larger fixed rational subgroups. |
| p | Count defective units and sample size varies. |
| np | Count defective units and sample size is fixed. |
| c | Count defects with constant opportunity. |
| u | Count defects with varying opportunity. |