Classification

Every map on this page draws the same values. What changes is the rule that turns numbers into colours, and it changes the map more than the data does. Quintiles make every map look equally varied; showing only the tails makes most of the city disappear; the top quarter answers “where does this concentrate?” and refuses every other question; a continuous ramp tells the truth about magnitude and hides the distinctions inside a crowded range.

This page offers eight rules, and the table under the map says what each one does and what it hides. None of them is wrong. They answer different questions, and a map that does not say which rule it used is not answering any of them.

Change the class rule

Class rule

Change city, level or measure (now New York, census tract, home value)

City

Geography

Form

Proportional symbols unavailable: 2,324 tracts of circles overlap into a blob at this level. Available at borough, PUMA and ZIP code.

Measure

Median home value across 2,324 tracts

Equal-count bins

ACS 2023 5-year, table B25077 · boundaries: US Census cartographic, 2020 vintage · suppressed estimates are shown as no-data, not as zero

Shows: Rank. Every map is fully saturated, so patterns are easy to see.
Hides: Magnitude. A city with a $10k spread and one with a $300k spread look identical.

The eight class rules

Each class rule this page offers, what it does, what it hides, and whether this level can carry it.
Rule	What it does	What it hides
Quantileson the map	Equal-count bins. Rank. Every map is fully saturated, so patterns are easy to see.	Magnitude. A city with a $10k spread and one with a $300k spread look identical.
Quantiles by population	Bins holding equal numbers of people. Where people are, rather than how many areas there are. Each shade covers an equal share of the population.	Comparability by number of areas. A shade may cover four block groups or four hundred, and the map cannot say which.
Equal interval	Bins of equal width. The shape of the distribution itself. A skewed measure piles into the lower bins, and that is the finding.	Nearly everything, on skewed data. A handful of high-income areas stretch the range, so median income piles most areas into the lower half of the bins and leaves the top ones nearly empty.
Natural breaks	Bins fitted to gaps in the data. Groups the data actually forms, by minimising variance inside each bin.	Comparability between maps. The breaks are fitted to this distribution, so two maps of the same measure in different cities share no scale.
Standard deviations	Distance from the mean. How unusual each area is, in units of the distribution's own spread.	The absolute values entirely, and it assumes a roughly symmetric distribution; on a skewed measure the mean is not the middle.
Top / bottom 5%	The extremes against everything else. Where the outliers are, which is usually the actual claim.	All variation inside the middle 90%.
Top 25%	Where the measure concentrates. Concentration. The quarter of areas with the highest values, against everything else.	All ranking inside the top quarter, and every difference among the other three quarters.
Continuous	Linear ramp across the range. True magnitude and skew. A long tail reads as a long tail.	Detail in crowded ranges. Most values pile into a few adjacent shades.