Quality Tools
Scatter Diagram Generator
Plot one variable against another to see whether — and how strongly — they are related.
Temperature vs Pressure Relationship
Legend
Correlation Analysis
Correlation (r): 0.9828
R-squared: 0.9659
Strength: Very Strong
Type: Positive
Linear Regression
Slope: 0.6945
Intercept: 13.7273
Equation: y = 0.69x + 13.73
Statistical Significance
Sample Size: 10
Significance: Significant (p < 0.05) · p < 0.0001
Variance Explained: 96.6%
📊 Correlation Interpretation
Correlation Strength: Very strong relationship between variables
Direction: Positive correlation - as X increases, Y tends to increase
Practical Application: Strong predictive model - suitable for forecasting
What is a scatter diagram?
Scatter diagrams (scatter plots) are graphical tools used to investigate the relationship between two variables and identify potential correlations. The independent variable goes on the X-axis, the dependent variable on the Y-axis, and each point is one pair of measurements.
Uses: process optimisation, testing suspected causes from a fishbone session, simple predictive models, and quality improvement.
Correlation types
| Type | Pattern | Example |
|---|---|---|
| Positive | As X increases, Y increases — upward trend | Temperature vs. ice cream sales |
| Negative | As X increases, Y decreases — downward trend | Price vs. demand |
| None | No apparent relationship — random scatter | Shoe size vs. intelligence |
Correlation strength guide
| |r| | Strength |
|---|---|
| 0.90 – 1.00 | Very strong |
| 0.70 – 0.89 | Strong |
| 0.50 – 0.69 | Moderate |
| 0.30 – 0.49 | Weak |
| 0.00 – 0.29 | Very weak |
- Correlation is not causation: a strong r does not prove one variable drives the other
- Outliers: a single extreme point can make or break a correlation
- Sample size: larger samples give more reliable estimates
- Non-linear relationships: r only measures straight-line association
Worked example: reaction temperature vs. yield
An illustrative chemical plant plots reaction temperature (X) against product yield (Y) for 30 batches over three months. Yield climbs with temperature at first, then flattens and falls at the hottest batches.
That curve is the point of the example: a straight-line correlation coefficient would understate the relationship, because the effect reverses past the peak. Looking at the plot — not just r — shows there is an optimum to operate at, which is then worth confirming with a designed experiment before changing the process.
Common pitfalls
| Pitfall | What to do |
|---|---|
| Spurious correlation | Two variables may move together by coincidence or because of a hidden third variable — check with process knowledge |
| Range restriction | A narrow data range can hide or weaken a real relationship — collect across the full operating range |
| Confounding variables | External factors may drive both variables — control or stratify for them |