Distortion Heatmap or Profile Report¶
Metric / dashboard — instantiates Mapping-Fidelity Distortion Control
A visualization or report showing where and how distortion varies by region, frequency, class, or operating condition.
Distortion Heatmap or Profile Report exists to defeat the average. Its whole reason for being is that a single fidelity number — "97% accurate," "0.4 units of error" — collapses a shape into a scalar and thereby hides exactly the distortions that matter: the corner that is badly wrong while the middle is fine, the frequency band that is smeared, the class that is systematically compressed. The mechanism takes the mapping's known transfer behavior and renders deviation across the operating space, so a viewer sees not how much distortion there is on average but where it lives and what pattern it makes. Its defining move is spatialization: distortion is laid out on the same coordinates as the thing being mapped, so a concentration jumps out visually instead of dissolving into a mean.
Example¶
A cartographer must choose a projection for a world map that will be used to compare the land areas of countries. Every flat map of a round Earth distorts something, and the danger is that a projection looks clean while grossly inflating high-latitude regions. Rather than trust a single "distortion score," the cartographer produces a distortion profile: at a grid of points across the map, small indicator ellipses are drawn whose size and shape encode how much the projection stretches area and skews angle at that location — Tissot's indicatrix. The map lights up: near the equator the ellipses are small circles (little distortion); toward the poles they balloon and elongate, revealing that Greenland's rendered area is several times its true relative size.
The report's payload is not a verdict but a picture: area is faithful in a broad equatorial band and increasingly unreliable poleward. Read against a stratified expectation — this map's job requires area fidelity within the populated mid-latitudes but tolerates polar exaggeration — the profile shows the projection is fit for purpose in the strata that matter and unusable in the strata that don't, which is precisely what an averaged number would have blurred away.
How it works¶
The report's craft is in what it makes visible:
- Derive from the transfer behavior. Distortion at each point comes from the mapping's known input→output behavior, so the picture reflects the mapping itself, not just sampled noise.
- Choose the coordinates. Lay deviation out along the dimension where concentration matters — spatial region, frequency, class, input magnitude, time.
- Encode magnitude and kind. Color, size, or glyph carry not only how much distortion but what type (stretch vs. shift vs. blur), so different failures read differently.
- Overlay the stratified expectation. Mark which regions must stay faithful and which may deviate, turning a descriptive picture into a fit-for-purpose read.
Tuning parameters¶
- Stratification axis — which dimension distortion is broken out along. The right axis reveals concentrated harm; the wrong one re-hides it under a different average.
- Resolution — how finely the space is gridded. Finer grids catch small hot spots but add noise and can invent structure that isn't there.
- Encoding scale — linear vs. perceptual color/size scaling. A poorly chosen scale can make a mild distortion look alarming or a severe one look benign.
- Aggregation window — how much data each cell pools. Wider windows are stabler but blur boundaries; narrower ones are sharp but jumpy.
When it helps, and when it misleads¶
Its strength is that it makes concentrated distortion impossible to miss and is the direct antidote to average-fidelity masking — the failure where a healthy headline mean coexists with a badly distorted sub-region.[n1] It also carries the mapping's transfer shape into the picture, so a viewer can reason about why a region distorts, and it gives the enforcement gate a stratified read to act on.
Its failure mode is the seduction of the picture itself: a heatmap with a well-chosen color scale can dramatize trivial variation or, worse, soothe with a smooth gradient that hides a sharp cliff at a poorly sampled boundary. It is descriptive, not decisive — it shows where distortion is but neither sets what is tolerable nor blocks anything, and a viewer can nod at a hot spot and ship anyway. The guarding discipline is to fix the stratification axis and color scale before looking at the data, sanity-check hot spots against raw cases rather than trusting the render, and pair the report with a gate that actually enforces the strata it reveals.
How it implements the components¶
distortion_profile— the report is the profile: deviation measured and laid out by region, frequency, or class rather than summarized to one number.stratified_distortion_budget— overlaying per-stratum expectations turns the raw picture into a fit-for-purpose read of which regions are within tolerance.mapping_transfer_model— the deviation shown is derived from the mapping's known input→output behavior, so the picture reflects the transform, not just sampled error.
It reveals distortion but does not block use when a stratum breaches its ceiling (distortion_budget_or_tolerance, appeal_or_override_channel) — that is Distortion-Budget Gate; nor does it reduce the deviation it displays (compensation_or_correction_rule) — that is Inverse Correction Mapping.
Related¶
- Instantiates: Mapping-Fidelity Distortion Control — makes the loop's distortion visible where it concentrates, so tolerance and correction can target it.
- Consumes: Transfer-Function Estimation supplies the transfer behavior the picture is derived from.
- Sibling mechanisms: Blind Reconstruction Comparison · Calibration Reference Set · Distortion-Budget Gate · Golden-Sample Regression Suite · Inverse Correction Mapping · Raw-Corrected Overlay Review · Residual Error Analysis · Transfer-Function Estimation
Editorial Notes¶
Form Classification¶
Form family: Interface, Display & Cue
Rationale: Distortion Heatmap or Profile Report operates as a user-facing prompt, display, template, or perceptual cue that shapes attention and action at the point of use because it a visualization or report showing where and how distortion varies by region, frequency, class, or operating condition.
Independent corroboration: The frozen evidence defines Distortion Heatmap or Profile Report as 'A visualization or report showing where and how distortion varies by region, frequency, class, or operating condition', so its operative form is Interface, Display & Cue.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Earth Sciences
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Cartographic projection practice cohered spatial distortion profiles, exemplified by Tissot's indicatrix showing where area and angle are altered.
Related originating lineages:
- Data Science & Analytics — Heatmaps and diagnostic visualization generalized local error profiles across classes and operating conditions.
- Mathematics — Geometry supplies the formal measures of mapping deformation being visualized.
Review resolution: Both current reviews place distortion_heatmap_or_profile_report primarily in earth_sciences; the reconciled classification retains only lineages that materially shaped the mechanism and keeps breadth of origin separate from reach.
Attribution caveat: The archetypal provenance is cartographic, while the generic heatmap/report form is modern data visualization.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
[n1] Tissot's indicatrix — Nicolas Auguste Tissot's device of drawing small ellipses across a map to show local distortion of area and angle introduced by a projection. It is the archetypal distortion profile: it displays where and how a mapping deviates rather than reducing fidelity to a single figure. ↩