Standards · Accountability

Data changelog

When a figure here turns out to be wrong, we fix it at the source and record it on this page. Below are the corrections we have applied to numbers that were already published, and when the EIA data behind the site was last refreshed.

Data refreshes

Read live from the pipeline's own record each time this page loads, so it cannot drift from the database serving the site.

Most recent pipeline rebuild
August 2, 2026
Every state, ranking and trend page was regenerated from the EIA tables in this build.
Price history held
2016–2025 (+ part of 2026)
10 complete years of EIA state price data, plus 2026 at about 38% of a full year as EIA publishes it. Counted from the database, not stated.

Corrections applied

Each entry says what a reader would have seen before the fix, not only what we changed. How corrections are handled, and how to report one, is on our editorial and corrections policy.

  1. Pages said "current" while the pipeline had been unable to run since April

    What was published: The database had last been rebuilt on 3 April 2026, and the freshness wording on the pages did not say so. The pipeline was not merely neglected, it could not run at all: the ETL declared its source handler at a path that resolved against the wrong directory, so downloads failed and the source reported as missing.

    What changed: The handler resolution and the three other breakages behind it were fixed, the pipeline was re-run, and freshness is now reported from the rebuild the database actually records.

  2. Renewable generation was defined inconsistently between pages

    What was published: Different pages counted the renewable share using different sets of fuels, so two pages could describe the same state with figures that did not agree.

    What changed: One definition is applied everywhere the renewable share is published.

  3. Guide pages carried hardcoded figures

    What was published: Several guide pages stated numbers typed into the page rather than read from the database, so they could not follow a data refresh and would silently go stale.

    What changed: Those figures now come from the same queries the rest of the site uses.

  4. Research pages carried a hidden "Reviewed by" claim

    What was published: A screen-reader-only line asserted that an editorial reviewer had checked the page on a given date. The visible page had always said "Compiled by", which is accurate; the hidden line claimed a review that never took place.

    What changed: The hidden claim was removed. We publish compiled data and say so; we do not claim a human review we did not perform.

  5. Seven chart data arrays were snapshots, not live queries

    What was published: Six guide pages and the electricity-cost calculator drew charts from arrays typed into the page. They matched the database at the time they were written, so nothing on screen was wrong, but they would not have followed the next refresh.

    What changed: All seven now read the database directly, so a chart cannot drift away from the figures beside it.

  6. The site claimed ten years of price history it did not hold

    What was published: /about and /methodology both said RateWatt retains and charts ten years of EIA price history. The database held nine distinct years, and the generation-mix series held eight.

    What changed: Both pages were corrected to the real window, and the note about the source dataset's own long publication history was separated from the span we actually ingest.

  7. A national trends line published a figure that was wrong by a factor of ten

    What was published: The lede described the 2024 residential rate as "roughly 17,500 hundredths-of-a-cent", which is $1.75/kWh, in the same sentence as the correct dynamic value of about 16.48¢/kWh. The wrong figure was typed in; the right one was computed.

    What changed: The hardcoded figures in that sentence were replaced with the values the page already derives.

What this log does and does not cover

This log begins on July 5, 2026, when we started recording corrections publicly. It is not a complete history of the site: changes made before that date are real but were not logged at the time, and we would rather say so than backfill entries we cannot stand behind.

It records corrections to published figures and refreshes of the underlying EIA data. It does not list routine design, layout or wording changes.

If a number here still looks wrong, tell us which page and which figure. That is how several of the entries above started.