Skip to content

Glossary

Analytics & BI

Heatmaps

Heatmaps are charts that use color intensity to show the size of a value across two dimensions, such as hour by weekday, or across a web page or map, so concentrations and gaps stand out at a glance.

Updated · By the VIZIO AI team

What is a heatmap?

"Heatmap" covers three different tools that share one idea: color stands in for magnitude.

  • Matrix heatmaps: a grid of rows and columns, such as a cohort retention table, a correlation matrix or orders by hour and weekday.

  • Website heatmaps: overlays on a page showing where visitors click, move or stop scrolling.

  • Density maps: geographic surfaces showing where points cluster, such as deliveries or incidents.

All of them trade precision for pattern: you see the hot zone instantly, but you cannot read exact values from color alone.

That makes a heatmap a scanning tool. It answers "where should I look?" well and "how big exactly?" poorly, so good versions print the value inside each cell or show it on hover.

How do you build a useful matrix heatmap?

Most of a heatmap's value comes from decisions made before coloring: the grid, the metric and the scale.

  1. Pick two dimensions with a manageable number of values, binning continuous ones (for example, hour of day).

  2. Choose a rate where totals mislead, such as conversion rate rather than conversions.

  3. Use a sequential palette for counts and a diverging palette when a midpoint matters, such as zero correlation or a target.

  4. Gray out cells with too few observations to trust.

  5. Sort rows and columns meaningfully, not alphabetically.

Seaborn, Tableau highlight tables and Power BI matrix conditional formatting all support this.

What do heatmaps look like in practice?

Website heatmaps are most useful for explaining why a page underperforms once analytics has shown that it does. Picture an online furniture retailer whose product pages get traffic but few add-to-cart clicks.

The team installs Microsoft Clarity and segments by device. The scroll map shows that, say, only a third of mobile visitors reach the delivery-time block near the bottom. The click map shows repeated taps on a lifestyle photo that is not a link. They move delivery information above the fold, make the photo open the gallery, and run an A/B test. The heatmap supplied the hypothesis; the test supplies the proof, which is how conversion rate optimization should work.

What are the limitations of heatmaps?

A heatmap is easy to misread because color hides sample size and exaggerates small differences. Watch for these traps:

  • Small samples: a page with 200 sessions produces confident-looking colors from noise.

  • Mixed devices: desktop and mobile layouts differ, so combined click maps are meaningless.

  • Raw counts on maps: density often just mirrors where people live.

  • Rainbow palettes, which create false boundaries between hues.

  • Privacy: session tools must mask form inputs, and in the EU they generally need cookie consent.

Dynamic pages add another problem: carousels, accordions and personalized content mean the page in the overlay may not match what each visitor saw. Check that your tool captures the rendered state, or segment by page variant before drawing conclusions.

How is a heatmap different from a choropleth map, and which tools are used?

A choropleth colors predefined areas such as states or postcodes; a density heatmap colors a continuous surface based on where points cluster. Choropleths suit rates per region; density maps suit raw event locations.

For websites, Hotjar, Microsoft Clarity and Contentsquare are common. For matrices and maps, Tableau, Power BI, Plotly, Seaborn, Kepler.gl and Mapbox are typical. See also data visualization for choosing between chart types and geospatial analytics for map-based work.

Matrix heatmaps also have a close relative, the cohort table, where rows are signup months and columns are months since signup. Colored by retention rate, it is one of the quickest ways to see whether newer customers stay longer than older ones.

Density heatmap vs choropleth map

Density heatmap

Choropleth map

Colors

A continuous surface

Predefined areas such as states

Input data

Individual point locations

Values aggregated per area

Best for

Where events cluster

Comparing rates between regions

Main pitfall

Mirrors population density

Large empty areas dominate visually

Example

Delivery drop-off points in a city

Conversion rate by state

How does VIZIO AI use heatmap-style views?

VIZIO AI's closest published work shows where activity concentrates across region, sector and time, which is the question heatmaps answer. For a UK cybersecurity research organization, Growth Studio built a Tableau dashboard showing attack counts by geographic distribution, affected industries and time-based trends, with drill-down filters by region, sector and incident type.

The case study does not specify heatmap charts; the overlap is concentration analysis at scale. Key views rendered in under one second on 500,000+ records, with a 25% reduction in mean time to detect. Where to start: pick the two dimensions your team argues about most, and plot a rate across them.

logo of Facebook
logo of X
logo of LinkedIn
logo of Reddit
logo of Telegram

Think AI-first.
Move future-fast.

Let’s build what’s next! From product ideas and intelligent workflows to data-driven systems and scalable growth operations.

Book a Strategy Call