Rising Demand, Weakening Supply: Persian Wikipedia as a Language-Technology Resource, 2016–2025, in 20 Language Editions

Purpose. Wikipedia is one of the main open text sources for training and evaluating language models, so the health of a language edition shapes the language technology available to its speakers. We ask whether Persian Wikipedia is under-supplied relative to demand and how its demand, editor base and content creation have evolved relative to 19 other languages. Design and methods. We combined Wikimedia statistics and monthly metrics for 20 language editions (user pageviews 2016–2025, active editors, new articles by registered users, anonymous users and bots) and compare Persian with the others in cross-section (supply, demand and editors, with regression-based prediction intervals) and in trajectories (percentage changes and ranks). Findings. In cross-section Persian is typical: 1.09 million articles, 2.24 billion user pageviews in 2025 (2,044 per article; peer median 1,664), 3,427 active editors, and 1.53 editors per million pageviews (peer median 1.47); observed editors were 0.95 of the number predicted from demand (95% prediction interval 0.49–1.86). Trajectories differ. Persian demand grew faster than in any other edition except Türkiye, whose edition was blocked until 2020 (+76.7% in 2019–2023 versus a peer median of +2.7%; first among 19 excluding Türkiye), then fell 24.0% in 2023–2025 (peers −13.8%). Active editors rose 26.4% in 2019–2021 and fell 27.2% in 2021–2025 (peers +7.8% and −16.9%), and fell 20.6% in 2025 alone (peers −9.8%; 18th of 20). Articles created by registered users peaked in 2022 (182% of the 2019 level against 105% for peers), fell by 53.4% in 2023 (peers −7.2%, the largest fall of 20) and were 16% below 2019 in 2025 (peers +7.6%). Bot-created articles, 45% of new articles in 2017, have been negligible since 2022. Originality. The study documents, with open metrics and a 20-language benchmark, that Persian Wikipedia's content creation and editor base weakened while demand rose, and separates the typical cross-sectional position of the edition from its atypical recent dynamics.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-10-18
DOI
https://doi.org/10.5281/zenodo.23448564
Primary Topic
Wikis in Education and Collaboration
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

Rising Demand, Weakening Supply: Persian Wikipedia as a Language-Technology Resource, 2016–2025, in 20 Language Editions

P. Ghorbani
Zenodo (CERN European Organization for Nuclear Research)
Wikis in Education and Collaboration
preprint

Rising Demand, Weakening Supply: Persian Wikipedia as a Language-Technology Resource, 2016–2025, in 20 Language Editions

P. Ghorbani
preprint en

Abstract

Purpose. Wikipedia is one of the main open text sources for training and evaluating language models, so the health of a language edition shapes the language technology available to its speakers. We ask whether Persian Wikipedia is under-supplied relative to demand and how its demand, editor base and content creation have evolved relative to 19 other languages. Design and methods. We combined Wikimedia statistics and monthly metrics for 20 language editions (user pageviews 2016–2025, active editors, new articles by registered users, anonymous users and bots) and compare Persian with the others in cross-section (supply, demand and editors, with regression-based prediction intervals) and in trajectories (percentage changes and ranks). Findings. In cross-section Persian is typical: 1.09 million articles, 2.24 billion user pageviews in 2025 (2,044 per article; peer median 1,664), 3,427 active editors, and 1.53 editors per million pageviews (peer median 1.47); observed editors were 0.95 of the number predicted from demand (95% prediction interval 0.49–1.86). Trajectories differ. Persian demand grew faster than in any other edition except Türkiye, whose edition was blocked until 2020 (+76.7% in 2019–2023 versus a peer median of +2.7%; first among 19 excluding Türkiye), then fell 24.0% in 2023–2025 (peers −13.8%). Active editors rose 26.4% in 2019–2021 and fell 27.2% in 2021–2025 (peers +7.8% and −16.9%), and fell 20.6% in 2025 alone (peers −9.8%; 18th of 20). Articles created by registered users peaked in 2022 (182% of the 2019 level against 105% for peers), fell by 53.4% in 2023 (peers −7.2%, the largest fall of 20) and were 16% below 2019 in 2025 (peers +7.6%). Bot-created articles, 45% of new articles in 2017, have been negligible since 2022. Originality. The study documents, with open metrics and a 20-language benchmark, that Persian Wikipedia's content creation and editor base weakened while demand rose, and separates the typical cross-sectional position of the edition from its atypical recent dynamics.

Zenodo (CERN European Organization for Nuclear Research)
Shahid Beheshti University (IR)
Wikis in Education and Collaboration
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.