Geographic information acceptance workshop / 6 MIN READ
Keep a map color change from masquerading as a business change
Accept the value-to-color contract before comparing regions or reporting a new hotspot.
AI-assisted research and analysis for the SNS Creative Fieldbook
Sources checked 2026-11-07
A choropleth map groups area values into colored classes. A change in class boundaries can recolor an unchanged value, so comparative maps need an explicit measurement and classification contract.
What you will leave with: A map-comparison specification with geographic units, value definition, class boundaries, missing-value treatment and controlled comparison fixtures.
Use this when
- A report compares colored territory maps across periods or business units.
- A dashboard automatically recalculates map classes whenever a filter changes.
Not a causal geographic experiment, an individual-customer inference, a territory allocation algorithm or a guarantee that a dark area needs more investment.
Define the mapped quantity and area first
Write the exact question the color should answer. Total enquiries, enquiries per eligible business and change in enquiries are different quantities. A large region can dominate a total while having a modest rate. Choose the quantity because it serves the question, not because it produces a more dramatic pattern. Keep any denominator’s date and coverage aligned with the numerator and geographic units.
Identify the boundary version and join key used to attach values to areas. An area split, merged district or mismatched label can change the picture before classification begins. Record missing, suppressed and genuinely zero values separately. A territory with unavailable evidence should not receive the same visual meaning as a measured zero simply because both were stored as empty cells during preparation.
Choose and disclose the class rule
Esri’s classification documentation explains that natural-break classes follow the particular dataset, while equal-interval and quantile methods answer different grouping questions. It warns that separately generated natural-break maps are unsuitable for direct comparison across different underlying data. The choice of class rule is therefore part of the analytical specification, not merely a styling preference.
Write the numeric boundaries and endpoint conventions alongside the legend. Specify what happens below, above and exactly on each boundary. For a time comparison, decide whether the question requires a common scale or relative rank within each period. Both can be useful, but a rank-colored series must not imply that the same shade represents the same absolute value. Make the intended interpretation visible in the report.
Use a controlled recoloring test
Create a synthetic region whose value remains fixed while values elsewhere change. Apply the proposed classification process to both periods and inspect its color. If the color changes, identify whether that reflects relative position or automatic boundary movement. The test does not prove one method is wrong; it reveals what a viewer may mistakenly infer from the visual comparison.
Add values immediately around a class boundary and a missing-value fixture. Confirm the legend, tooltips and exported image describe the same classes. A map can preserve its numbers while a rounded legend hides which side of a threshold a region occupies. Keep a small accompanying table with the actual values so reviewers can separate the measured change from the display’s grouping choices.
Publish the comparison with its limits
Retain the mapped field, denominator when used, boundary version, class method and explicit breaks with the report artifact. When two maps use different rules, label that difference prominently or redesign the comparison. Do not silently hold a legend fixed while allowing the colors to recalculate, and do not quietly recompute earlier maps without recording the revision.
Ask reviewers to explain what one dark region and one newly recolored region mean before approving the narrative. If they infer a larger absolute change than the data supports, repair the presentation or wording. Area-level patterns do not describe every customer within an area, and visual proximity does not establish a campaign effect. The accepted artifact demonstrates a consistent value-to-color contract, not an investment recommendation or proof of causality.
Make the reasoning concrete.
A fictional trade publisher maps enquiries per eligible business by sales territory each quarter. One territory has exactly the same rate in both quarters, yet its color darkens because the dashboard recalculates natural-break classes from the new overall distribution.
Describe a new hotspot
Opening: Write that the territory’s interest increased because the shade darkened.
Hypothesis: The visual appears to support a simple growth story.
Risk: The underlying territory value did not increase.
Inspect the class contract
Opening: Compare numeric values and breaks, then choose a common scale or explicitly relative view.
Hypothesis: The report tells readers what changed in the data and what changed in the grouping.
Risk: A common scale may be less visually dramatic and still be the right comparison.
The publisher uses common breaks for the quarter-to-quarter value comparison and preserves a separately labeled relative-position view if needed. It removes the unsupported growth claim rather than treating a color shift as an observed business outcome.
Take the questions into your next brief.
Use these prompts in your own notes. Nothing is submitted from this page.
Measurement
What quantity, denominator and reference period does each area encode?
Geography
Which boundary version and join key define the areas?
Classes
What rule, explicit breaks and endpoint conventions map values to colors?
Control
Does an unchanged synthetic region change shade when other values move?
Interpretation
Can a reader distinguish absolute change, relative position and missing evidence?
Watch for the wrong conclusion.
- Treating a darker shade as evidence that the area’s value increased.
- Comparing independently recalculated legends as if they shared a scale.
- Giving missing evidence the same meaning as a measured zero.
Your next action
Run an unchanged-region fixture through two periods of your proposed map classification before writing the comparison narrative.
Sources & scope
Workflows, fixtures, worksheets and hypothetical examples are original editorial analysis. Sources support the specifically attributed technical claims. No real location data is transmitted, API enabled, account changed or geographic decision implemented.
- Data classification methods →Primary platform documentation · Supports: Supports distinctions among natural breaks, equal interval and quantile classification, including the comparison limitation of data-specific natural breaks. Comparison fixtures and editorial workflow are original.