How to Make a Scatter-Plot Matrix on Mac & Windows

A scatter-plot matrix shows every pair of variables at once — the “see everything” view. Here's how to make one on Mac & Windows and use it to spot relationships before you model.

A scatter-plot matrix (SPLOM) is a grid of small scatter plots, one for every pair of variables. It's the classic “show me everything at once” chart for spotting which pairs relate before you commit to a model — and Stratum draws a real one.

Make it

Open your dataset and choose the Scatter chart.
Select several numeric variables.
Turn on Matrix mode — Stratum lays out a grid of scatter plots, every pair against every other, with shared axes.

A scatter-plot matrix in Stratum's Scatter chart Matrix mode — Matrix mode: every variable pair, side by side.

Read the grid

Each off-diagonal cell is one pair's scatter. Scan for cells with a clear upward or downward trend (a strong relationship), a curve (non-linear), or a stray point (an outlier). Because every pair is on screen at once, patterns — and the variables that drive them — jump out faster than flipping through plots one at a time.

Add the numbers

The matrix shows you the shape of each pair; for the exact strength, pair it with a correlation matrix, which puts a Pearson or Spearman coefficient on every pair and highlights the significant ones.

A correlation matrix alongside the scatter-plot matrix in Stratum — The correlation matrix adds the exact strength of each pair.

Scan, then confirm. The scatter-plot matrix is the shape view; the correlation matrix is the strength view. Together they tell you which relationships are real, and which are linear — the whole point of exploring before you model.

Why this matters before modeling

Seeing the pairwise structure first saves you from surprises: redundant predictors (multicollinearity), non-linear relationships a straight model would miss, and outliers that distort everything. See Lesson 12.

Download Stratum

Frequently asked questions

What is a scatter-plot matrix (SPLOM)?

A grid of scatter plots showing every pair of variables at once, used to spot which pairs relate before modeling. In Stratum, turn on Matrix mode on the Scatter chart to draw one.

How do I see all my variables' relationships at once on Mac & Windows?

Use the Scatter chart's Matrix mode for a grid of every pair, and a correlation matrix for the exact strength of each relationship.

Why look at pairwise relationships before regression?

To catch multicollinearity (redundant predictors), non-linear relationships, and outliers up front — issues that quietly distort a model if you skip the exploration.

Scatter-plot matrix or correlation matrix — which do I need?

Both. The scatter-plot matrix shows the shape of each pair; the correlation matrix shows the strength as a number. Scan shapes in the SPLOM, read strengths in the matrix.