A graduated symbol map displays quantitative data by grouping values into a small number of classes, then assigning each class a symbol of a specific size. If you’re mapping city populations across a country, for example, cities with 0–100,000 people might get a small circle, cities with 100,001–500,000 a medium circle, and cities over 500,000 a large circle. The key feature is that the data gets sorted into discrete bins rather than scaled continuously, making the map easier to read at a glance.
How Graduated Symbols Work
The process starts with a dataset of raw numbers tied to geographic locations: population counts, earthquake magnitudes, hospital capacities, tons of grain produced. The mapmaker divides those numbers into classes (typically three to seven), and each class gets its own symbol size. Every data point within a given range is represented by the same symbol, regardless of where it falls within that range.
This is what cartographers call “range scaling” or “range grading.” It simplifies the visual picture significantly. A reader can look at the legend, see three or four distinct circle sizes, and quickly match each symbol on the map to its class. The tradeoff is that exact values are lost. A city of 120,000 and a city of 480,000 might both land in the same class and look identical on the map.
Graduated vs. Proportional Symbol Maps
This distinction trips up a lot of people because the two map types look similar. A proportional symbol map uses “absolute scaling,” meaning each symbol’s size is mathematically proportional to the actual data value it represents. A city twice as populous gets a symbol with twice the area. Every data point can have a unique symbol size, so you might see dozens of slightly different circles on a single map.
A graduated symbol map, by contrast, groups values into classes first. You’ll only see a handful of distinct symbol sizes. In GIS software like ArcGIS, these are treated as separate map types: selecting “proportional symbols” produces continuously scaled symbols, while selecting “graduated symbols” produces range-graded ones.
Proportional maps preserve more precision but are harder to read. People consistently underestimate differences in circle area, so two symbols that differ by a factor of two don’t look that way to most viewers. Graduated maps sidestep this problem by reducing the comparison to a few clearly distinct sizes.
Choosing Classification Methods
How you divide your data into classes changes the story the map tells. The most common methods are:
- Equal interval: Splits the full data range into classes of equal width. Simple to understand, but if your data clusters at one end, most symbols may fall into a single class.
- Quantile: Places an equal number of data points in each class. This spreads symbols evenly across the map but can group very different values together or split similar values apart.
- Natural breaks (Jenks): Uses an algorithm to find natural groupings in the data, minimizing variation within each class and maximizing variation between classes. This usually produces the most intuitive-looking map.
- Mean-standard deviation: Centers classes around the data’s average. Useful when you want to highlight how far values fall above or below the mean.
For most general-purpose maps, natural breaks is the default choice in GIS software because it respects the structure of the data. But if your audience needs to compare across multiple maps or time periods, equal interval keeps the class boundaries consistent.
When To Use a Graduated Symbol Map
Graduated symbol maps are best suited for count data: total population, number of hospitals, reported cases of a disease, units sold. These are raw totals tied to specific locations or areas. If you tried to show count data on a choropleth map (the kind that shades entire regions in different colors), the result would be misleading because large geographic areas would dominate the visual impression regardless of their actual values. A huge, sparsely populated state would look more significant than a tiny, densely packed one.
If your data is a rate or ratio, like people per square mile or percentage below the poverty line, a choropleth map is typically the better fit. The general rule: counts go on symbol maps, rates go on choropleth maps. If you want to show counts on a choropleth, you need to standardize the data first (convert it to a rate).
Graduated symbols also work well for three-dimensional symbols like prisms or spheres, where continuous proportional scaling makes size differences even harder to judge by eye.
Why Readers Misperceive Circle Sizes
Human vision is not great at comparing areas. When people look at two circles and one has twice the area of the other, most viewers perceive the difference as smaller than it actually is. Research dating back to the 1950s by cartographers James Flannery and Robert Williams documented this systematically, finding that people consistently underestimate larger symbols relative to smaller ones.
Flannery developed a mathematical correction factor that slightly inflates larger symbols to compensate for this perceptual bias. Many GIS tools apply this correction automatically when creating proportional symbol maps. Graduated symbol maps largely avoid the problem altogether: because there are only a few distinct sizes and readers match them to a legend rather than estimating ratios, the perceptual error matters much less.
Dealing With Overlapping Symbols
When data points are close together, their symbols can pile up and obscure each other. This is one of the practical challenges of any symbol map, and it gets worse as symbol sizes increase. A few techniques help:
Transparency lets you see symbols stacked behind others. Setting symbols to 30–50% opacity means overlapping areas appear darker, which actually adds information by showing where data points cluster. Symbol ordering, where smaller symbols are drawn on top of larger ones, prevents big circles from hiding small ones entirely.
GIS software also offers displacement tools. ArcGIS Pro’s Disperse Markers tool, for instance, detects overlapping point symbols and nudges them apart based on a minimum spacing you set. This preserves readability at the cost of slight geographic imprecision. For dense datasets, you may need to reduce the number of classes or shrink the overall symbol size range to keep the map legible.
Designing a Clear Legend
The legend is what makes a graduated symbol map interpretable. Each class needs a symbol shown at its exact display size, paired with the value range it represents. Nesting the symbols (stacking them concentrically so the smallest sits inside the largest) saves space and makes size comparisons more intuitive than listing them separately.
Label each class with its full range, not just its midpoint. A reader seeing a medium circle should be able to check the legend and learn it represents “100,001 to 500,000” rather than just “300,000.” Round your class boundaries to clean numbers when possible. A break at 500,000 is easier to process than one at 487,312, and the slight loss of statistical precision rarely matters for the reader’s understanding.