Edit Forensics · English Wikipedia · 2020–2024

TrangaBellam

A neutral, data-driven dossier of one editor's complete public record — 22,364 edits across 3,370 pages, pulled straight from the MediaWiki API and independently fact-checked.

USER ID 39729082 REGISTERED 2020-07-13 LAST EDIT 2024-11-04 BLOCK LOG none RIGHTS reviewer · rollbacker · autoreviewer
0
Edits captured
0
Pages touched
0
Articles created
0
Years active
scroll ↓
01 — Where the work happens

Half builder, half debater.

Only about 48% of edits land in articles. The majority live in discussion and process space — talk pages, user talk, and Wikipedia's dispute-resolution noticeboards. This is a discussion-heavy profile, typical of an editor working in contentious topic areas.

The discussion tilt

52%
of all edits are in Talk, User-talk and Wikipedia-process namespaces — not articles.
9.3%
sit in Wikipedia-space alone, concentrated in the machinery of disputes (AfD, ANI, RSN, AE).
02 — Four years, monthly

A sharp ramp, a long taper.

Activity explodes through 2021 (8,540 edits) and stays hot through 2022 (7,746), then tapers steadily. The single near-silent month — April 2023, just 7 edits — was a personal break, not a block: the account has a clean block log and was editing arbitration pages that same month.

03 — The center of gravity

Pre-modern history meets contested present.

Each row carries two shares: of edits (how many of the page's revisions this account made) and of content (how much of the text live on the page today they actually wrote, per the WikiWho authorship engine). The gap is the story — on the history articles the surviving-text share dwarfs the edit share (deep authorship: 87.9% of Multan Sun Temple, 86.4% of Gyanvapi), while on The Kashmir Files both stay low. Open any page for its full breakdown.

share of edits share of surviving contentclick a row →
04 — Builds and prunes

Adds almost exactly what it removes.

Across four years the net byte change is nearly flat — this is not a one-directional "adder" or a pure "deleter." The largest single moves are mostly reverts and restorations on contested articles (Sati, Talk:Sun), not greenfield prose.

Net change across 22,364 edits: +122,512 bytes — a rounding error on 6.6 million added and 6.5 million removed.
05 — The dispute machinery

Fluent in process.

A strongly enforcement-oriented footprint: hundreds of edits to deletion debates and the noticeboards where contentious-topic conduct is litigated.

06 — Tone & bias — signals, not a verdict

What the summaries suggest.

Analysis here rests on edit summaries, page titles, namespaces and byte deltas — not the diff content itself. Findings are signals for human review. Every large move inspected carried a stated policy or talk-page rationale, so the record is consistent with both principled sourcing enforcement and a directional slant. Distinguishing the two requires reading the diffs.

Tone

Terse baseline, blunt streak

Mostly telegraphic summaries ("ce", "+", "rm"), punctuated by a recurring sardonic-combative register.

verbatim: "Nonsense" · "Utter nonsense" · "Hindutva propaganda" · "fringe racist nonsense"
Collegiality

Real, alongside the edge

The same record shows genuine good-faith collaboration — thanks given, assume-good-faith reverts, sourcing help on the Resource Exchange.

71 "thanks" · AGF-tagged reverts · 175 edits to Resource Request
Directional signal

Net removal on grievance pages

Net byte-removal on Hindu-grievance / Hindutva-aligned pages; net addition on academic critics of Hindutva. A signal to verify against diffs — with policy rationales attached throughout.

Persecution of Hindus, The Kashmir Files, HAF ↓ · Audrey Truschke ↑
Conflict locus

~11% article-edit revert rate

Reverts/undos cluster tightly in South-Asian history and Hindu/Hindutva-adjacent BLPs, mirrored by heavy talk-page presence on the same titles.

~1,177 of 10,685 article edits are reverts
The data is consistent with both principled NPOV/RS enforcement and selective slant — and the summaries alone cannot tell them apart. — Fact-checked conclusion, tone & bias analysis
07 — Integrity & method

Checked against the raw data.

Two independent fact-checkers re-derived every quantitative claim from the source data. Verdict on both reports: mostly sound. Corrections below are already reflected on this page.

  • Namespace counts, byte flow (+6,596,510 / −6,473,998), page creations (328), and the central bias table verified exactly against raw API data.
  • All spot-checked edit-summary quotes confirmed genuine and correctly attributed to their revision IDs.
  • Every largest-addition / largest-removal revision verified to the byte.
  • Corrected: the April-2023 lull was not a block — no block log exists; account was active on arbitration pages that month.
  • Corrected downward: "sharp-removal" phrasing (~7, not 38), AGF-tagged reverts (~10, not 24), GA/DYK summaries (~246, not 443).
  • Activity-type and topic-domain buckets are heuristic groupings, not exact API fields — read as approximate.
Scope & fairness. This is an independent data-analysis exercise built entirely from public MediaWiki contribution logs. It does not pull diff content, outcome data, or off-wiki context. Off-wiki allegations about this editor exist in partisan press; they are not established facts and are not asserted here. Nothing on this page is a finding of misconduct — the editor has a clean block log and trusted-editor permissions. Treat every "bias signal" as a prompt for diff-level review, not a conclusion.