Data Changelog
This page lists confirmed changes to PlainClimate data handling, methodology, and accountability surfaces. We record only changes that can be checked against a commit, a pipeline rebuild, or a documented editorial action. The same entries also appear under editorial policy.
August 9, 2026
- Accountability routes. Added this dedicated
/changelogpage and a permanent/correctionsalias to the editorial corrections process, so typed or impressed accountability URLs resolve instead of soft-404ing.
August 4, 2026
- Score explainer honesty. Score explainers that incorrectly named humidity and extreme-weather frequency as inputs were aligned with the implemented four-field formula. City pages now expose each component, its points, missing-data normalization, and corpus-derived grade thresholds.
July 23, 2026
- Full NOAA rebuild + comfort-score N/A rule. Cities with fewer than two of the three optional components (precipitation, snowfall, growing-season length) now show N/A instead of a silently imputed midpoint value.
- Retired unsupported tool. Removed the warming-anomaly lookup estimator: 30-year normals cannot support per-city current-anomaly estimates.
July 21–22, 2026
- Histogram + display honesty. National comfort-score histogram no longer drops perfect-100 cities from the top bucket; percentiles capped at 99%; frost/snowfall copy corrected where source data is zero or missing.
Reporting a possible problem
Use the contact page with the URL and the value that looks wrong. The four-step fix-at-source process is on editorial policy.
Open data: plainclimate-city-normals.csv (NOAA 1991–2020 U.S. Climate Normals · public domain).
Changelog entries are limited to corroborated pipeline or git events; empty honesty beats invented history. No figure on this page is typed in by an editor - each value is rendered from NOAA 1991-2020 U.S. Climate Normals. See our editorial standards & corrections policy, the methodology behind these numbers, or report a data error.