What Is a Matrix Heatmap?
A matrix heatmap is a grid where two categories — one on each axis — intersect in cells, and each cell is colored based on its value. Instead of reading numbers row by row, you see the pattern instantly: darker or brighter cells jump out, showing you where the concentration, the outliers, or the gaps are.
It's the same idea as a spreadsheet's conditional formatting, but built for exploration rather than static reporting — you can usually click, filter, and drill into any cell.
When to Use a Matrix Heatmap
A heatmap earns its place when you're comparing two categorical dimensions at once and want to spot patterns, not just totals. Common cases:
- Coverage analysis — which combinations of two categories (e.g. service level × region) have data, and which are empty
- Concentration analysis — where the bulk of a metric sits (e.g. which channel and quarter drive most revenue)
- Quality or performance scoring — average scores across two dimensions, to spot weak intersections at a glance
- Anomaly spotting — a single unusually high or low cell that would be easy to miss in a table
If you only have one category to analyze, a bar chart is usually clearer. Heatmaps shine specifically when there are two dimensions interacting.
A Simple Example
Imagine you're looking at customer orders and want to understand delivery performance. You have two categories — picking type (how an order was picked) and delivery distance (how far it traveled) — and you want to know how many distinct customers fall into each combination.
Laid out as a plain table, this might be dozens of rows: one per picking type, repeated for every distance bracket. Scanning it for patterns means mentally cross-referencing numbers across the whole table.
Turn the same data into a heatmap — picking type on one axis, delivery distance on the other, customer count coloring each cell — and the pattern appears instantly. Maybe most customers cluster in "standard picking, short distance," while "express picking, long distance" is nearly empty. That's a five-second read instead of a five-minute one.
Heatmap vs. Other Chart Types
It helps to know when a heatmap is the right tool versus the alternatives:
- Table with conditional formatting — works for small datasets, but doesn't scale past a handful of rows/columns before it becomes visual noise.
- Bar or column chart — better when you're comparing one dimension, not two. Stacking a second dimension into a bar chart quickly gets cluttered.
- Scatter plot — better for two continuous numeric variables (e.g. price vs. quantity), not two categorical ones.
- Matrix heatmap — the right choice specifically when both axes are categories and you want to compare a metric across every combination of the two at once.
What It Takes to Build One
A basic heatmap needs three things: two categorical fields (one per axis) and a numeric value to color the cells by. The tricky part in most BI tools isn't the chart itself — it's getting the underlying numbers right.
Two situations trip people up constantly:
Counting unique things across two dimensions. If you want to know how many distinct customers fall into each row/column combination, a normal sum or count will overcount — a customer who appears under two categories gets counted twice when you try to recombine measures into a grid. This needs to be computed directly from the raw ID column, not a pre-aggregated measure.
Totals that don't add up the way you'd expect. When a heatmap shows distinct counts, the row and column totals are genuine distinct counts across the whole row — not the sum of the individual cells. That's mathematically correct, but it looks "wrong" at first glance if you're expecting simple addition.
How Dynamic Matrix Heatmap Handles This
Dynamic Matrix Heatmap is a Power BI custom visual built around exactly this kind of analysis — with the in-canvas flexibility to change what you're looking at without leaving the report.
- Pick your axes on the fly. Drag fields into Dimensions and Measures, then choose X, Y, and Value directly inside the visual — no need to rebuild the chart to explore a different angle.
- Exact distinct counts, done right. Drop a raw ID column into Distinct Count Of and the visual computes true unique counts per cell — solving the double-counting problem above without any DAX gymnastics.
- Show value as % of row, column, or grand total — useful for answering "where does this data actually concentrate?" instead of just "what's the number."
- Totals that are honest. Row and column totals reflect real distinct counts, with a clear explanation of why they may be smaller than the sum of visible cells.
- Scales from simple to complex. Free covers up to 3 dimensions and 2 measures — enough for most reports. Pro extends that to 10 dimensions, 5 measures, and unlocks distinct counts, percentage views, and totals.
Nothing is watermarked or time-limited — Free is a genuinely usable starting point, not a trial.
Try It
If you're working with two-dimensional categorical data and need more than a plain table can show you, Dynamic Matrix Heatmap is built for exactly that.