Basics: What Is a Heat Map? Why the Legend Matters
What is a heat map? A heat map uses color to represent numerical values: in a grid of store orders, one dark Wednesday cell can reveal a rush that disappears in a page of figures. Each colored cell or area ties a measurement to a position, whether that is a calendar day, a geographic location, or a webpage button.
That dark patch is a starting point. Reading it means connecting its position to its value, then asking what produced the difference.
In This Article
- Read the Axes and Legend First
- Recognize the Main Types
- Choose Colors That Match the Values
- Arrange Your Data Before Mapping It
- Build a Small Example in Excel
- Check Patterns Before Drawing Conclusions
- Key Takeaways
- Frequently Asked Questions
- Start With One Clear Comparison
Read the Axes and Legend First
Suppose a store manager sees a dark cell labeled “Sales” and schedules extra packing staff, assuming it means more orders. If the unit is dollars, one expensive purchase could explain the patch. The missing unit has turned revenue into an imaginary workload.
Rows and columns tell you where to look; the legend tells you what you found. Check its endpoints and direction: darker or brighter can mean higher, depending on the scale. Check whether colors change continuously or in bands, too. In a band labeled “50–99,” a store with 50 visitors looks identical to one with 99. Almost twice the traffic. Exactly the same color.
Recognize the Main Types
A matrix compares values at row-column intersections, such as support requests by weekday and hour. Geographic heat maps estimate a density surface from point data. Choropleth maps instead color predefined geographic areas by their values; they are a related visualization, not the same geographic method.
Website maps record behavior. One revealing case is a click hotspot on something that does nothing: in an imagined shop page, visitors repeatedly click a product photograph expecting it to open. The patch could prompt a design fix, not a celebration of engagement. Scroll and pointer maps answer different questions, as the table shows.
| Heat map type or related map | What color measures | Best for | Main advantage | Key limitation |
|---|---|---|---|---|
| Matrix heat map | Cell value, such as monthly revenue ($) by product | Products across months | Shows highs and lows across two categories | Exact values need labels or tooltips |
| Correlation heat map | Pairwise correlation coefficient; Pearson r ranges from −1 to +1, with 0 at the midpoint | Linear relationships among numeric variables | Summarizes variable pairs | Correlation does not establish causation |
| Geographic density heat map | Point concentration or weighted density within a chosen radius | Clusters of incidents, stores, or trips | Shows hotspots without administrative boundaries | Radius or bandwidth changes hotspot size and intensity |
| Choropleth map (related map, not a geographic heat map) | Regional value, such as residents per km² | Rates across counties, states, or countries | Uses meaningful boundaries | Large regions dominate; local differences disappear |
| Website click heat map | Recorded click count or share by page location | Popular links and noninteractive elements | Shows attempted interactions | Clicks can mean interest or frustration |
| Website scroll heat map | Percentage of recorded visits reaching each depth | Content or calls to action | Shows declining exposure down a page | Reaching content does not prove reading |
| Website pointer-movement heat map | Recorded cursor positions or movement density | Desktop mouse-use patterns | Shows activity beyond clicks | Not eye tracking; touch-only visits lack mouse-hover paths, but finger movement can be tracked |
Choose Colors That Match the Values
For order counts, a sequential palette such as light-to-dark blue communicates increasing amounts. For gains and losses around zero, use a diverging palette with contrasting ends and a neutral midpoint. That midpoint carries meaning; it is not decoration.
Color can also smuggle in a judgment. Consider a hypothetical executive review where red marks the highest revenue, but a director reads it as a warning and cuts stock at the busiest store. Red-green confusion adds another risk for readers who cannot readily distinguish those hues. Use clear lightness differences and explain whether a color means high, low, favorable, or unfavorable. In a temperature chart, red may simply mean warmer. Keep cell values visible. Shades are not numbers.
Arrange Your Data Before Mapping It
Keep one measure and consistent units in a matrix: revenue in dollars and conversion rates do not belong on the same unadjusted scale. Preserve calendar order for months; for unordered categories, choose the arrangement that serves your comparison.
| Data arrangement | Best for | Pros | Cons |
|---|---|---|---|
| Chronological order | Days, months, quarters | Shows seasonality and streaks | Similar periods may be distant; unequal intervals need labels |
| Natural numeric order | Ages, temperatures, scores | Preserves progression: 18–24, 25–34, 35–44 | Unequal band widths can mislead |
| Alphabetical order | Stores, products, departments | Makes labels easy to find | Neighbors may be unrelated |
| Ranked by row total | Store purchase totals | Groups high-volume stores | Can hide small-store peaks; period affects ranking |
| Grouped by category | Stores within regions; products within departments | Supports within-group comparison | Boundaries can obscure similarities |
| Clustered by value pattern | Similar weekly store patterns | Places similar rows together | Depends on scaling, distance metric, and clustering method; order may change |
Ordering brings cells together; missing data can create a false connection between them. A store with no purchases deserves a zero. A store whose reporting system failed needs a distinct neutral fill or “No data” marker explained in the legend. Turning both into zero makes a reporting failure look like a business failure.
Build a Small Example in Excel
Use an imagined bread shop taking preorders during a Wednesday promotion. In Excel for Microsoft 365 on Windows, paste this dataset into A1. All values are orders per day.
| Store | Monday | Tuesday | Wednesday | Thursday | Friday |
|---|---|---|---|---|---|
| Central | 12 | 18 | 41 | 20 | 15 |
| Riverside | 10 | 14 | 35 | 17 | 13 |
| Market | 8 | 11 | 29 | 14 | 10 |
Select B2:F4, leaving the headings out. Choose Home > Conditional Formatting > Color Scales and select a two-color scale. Then choose Home > Conditional Formatting > Manage Rules, select the rule and click Edit Rule. Set Format Style to 2-Color Scale. For both endpoints, set Type to Number: use a minimum of 0 with a white fill and a maximum of 50 with a blue fill. Confirm both dialogs. Keep the numbers visible.
For a matching legend, enter “Orders per day” in H1 and 0, 10, 20, 30 and 50 in H2:H6. Apply the same two-color rule, endpoint colors and fixed bounds to H2:H6. Use these bounds for any equivalent grid you want to compare.
Wednesday forms a darker vertical band: every store has its highest count that day. Central’s 41 orders is the grid’s maximum, but not the scale’s upper bound of 50. The promotion is a possible explanation for the shared peak, not something the colors prove.
Check Patterns Before Drawing Conclusions
Suppose the bread shop also stayed open longer on Wednesday. Its higher total now combines extra selling time with the promotion. Orders per opening hour would answer a different question about demand. Geographic incident counts have a similar denominator problem: a busy population center may have more incidents without a higher rate per resident.
Before acting on an unexpected patch, inspect the underlying records. For website maps, compare the same page version and account for mobile versus desktop layouts; the same button may occupy different positions.
Across charts, keep units and scale limits compatible. Otherwise, each chart’s darkest cell can represent a different quantity. Matching colors on mismatched scales create a false comparison.
Key Takeaways
- Treat a hotspot as a question worth investigating, not an explanation.
- Choose the map type for the measurement, not its appearance.
- Make units and color meaning explicit.
- Keep missing data visibly separate from zero.
- Compare equivalent records on compatible scales before making a decision.
Frequently Asked Questions
Heat maps also turn up in floor plans and warehouse operations. These applications use the same basic encoding, but the measurements differ.
Can you give me some examples of heatmaps?
Examples include a calendar colored by daily exercise minutes and a warehouse grid colored by visits to each storage location. The calendar can reveal routines and gaps, while the warehouse example highlights frequently used areas.
What is another name for a heatmap?
A matrix heat map is also called a color-coded matrix or shaded matrix. “Heatmap” and “heat map” are alternative spellings; names such as correlation matrix and density map refer to particular forms, not interchangeable names for every version.
How do I create a heat map?
Follow the Excel example above, then check the result: Central’s Wednesday cell should show 41 and be darker than its Monday cell showing 12.
What is a Wi-Fi heat map?
A Wi-Fi heat map overlays wireless measurements, such as signal strength, on a floor plan or geographic area to reveal coverage gaps. Survey-based maps use measurements, while predictive maps estimate coverage from access-point placement and building characteristics.
How much does a heat map cost?
The spreadsheet example needs no separate heat-map service if you already have Excel. For website analytics or a Wi-Fi survey, first specify which service you need; they are different purchases.
Start With One Clear Comparison
Test the finished bread-shop heat map with someone who has not read these instructions. Ask: “When does each store receive the most orders, and how many does Central receive that day?” They should identify Wednesday and 41 orders. If they cannot, check the labels, legend and number visibility before adding more data.
Recommended for you
Loanable Funds Graph: Cheaper Credit, Less Investment
A loanable funds graph shows why weaker investment demand can lower interest rates. Distinguish curve shifts from movements and trace the cause. See how.
Design Principles for Beautiful and Readable Network Maps
Put 2,000 nodes and 8,000 edges on a canvas and most viewers will see spaghetti. Use layout constraints, color discipline, and geometry tuned to human perception, and clusters, bridges, and outliers become legible in under five seconds—the difference between guesswork and insight. This guide distills Design Principles for Beautiful and Readable Network Maps into practical […]
Math Line Plot: Each Mark a Measurement, Each Stack a Count
A line plot turns nine bean pod lengths into visible counts. Learn what marks and stacks mean, then build your own with fractions. See the worked example.
Sin Graph From the Unit Circle: Five Points, One Wave
Build a sin graph from unit-circle heights and five key points. Learn amplitude, period, and shifts, then read equations from curves. See the steps inside.
What Is an Infographic? How Visuals Earn Their Place
What is an infographic? See how an electricity chart and appliance diagram can work together to explain a message, not just decorate it. Explore the basics.
What Is a Diagram? Explanation, Not Just Illustration
What is a diagram? Learn how labeled shapes and connecting lines explain relationships, how diagrams differ from drawings, and how to read one. Read on.
Money Market Graph: The Cash Squeeze Behind Rate Rises
A money market graph links cash needs to interest rates. Learn how to read its axes, find equilibrium and track supply or demand shifts. See the breakdown.
Circular Flow Diagram: A $4 Loaf Links Two Markets
A circular flow diagram connects a $4 bread purchase to work and income. Learn how goods and money move through two markets. Follow the example in the article.
