Trending Wikipedia articles (weekly)

Weekly ranking of what the English-speaking world is reading on Wikipedia: every article that appeared in the daily top-1000 (all access types) over the trailing 7 complete days, merged on the canonical title, from the official keyless Wikimedia pageviews API (https://wikimedia.org/api/rest_v1/, CC0 1.0). Deterministic pipeline: namespace pages excluded (Main Page, Special:, Category:, Wikipedia:, File:, Template:, Portal:, Draft:, ...); cross-day de-duplication on the canonical title; each article scored for attention (0-100 min-max of log1p(total_views)) and growth (trend_pct = mean views of the second half of the window minus mean views of the first half, over the first-half mean), combined as trend_score = 100 * (0.6 * attention + 0.4 * growth_mapped) where growth_mapped = clip((trend_pct + 1) / 4, 0, 1), and ranked (trend_score desc, total_views desc, title asc) as trend_rank. breakout flags articles whose views at least doubled vs the first half of the week on sustained second-half attention (>= 1000 mean views/day). trend_pct is capped at 99.0 (values at the cap mean negligible first-half presence). topic is a deterministic first-match-wins rule set over the article's MediaWiki categories (sports, politics, entertainment, science, technology, business, biography, history, geography, culture, other); topic_categories records the categories used. Week-granular as-of stamping (fetched_at = window-end midnight UTC): a second ingest of the same window is a true no-op. Primary key: (week, article_title). Nullability: no nulls are emitted. Caveats: topic is a keyword rule set over categories, not a classifier; the top-1000/day cutoff means articles with modest but steady readership never appear; English Wikipedia only. Underlying aggregates are CC0 1.0, so commercial_use = yes. Sample use: order by trend_rank for the fastest-rising articles of the week, or filter topic = 'politics'. Note: this is a weekly cross-topic leaderboard of the most-viewed articles; it complements wiki_attention_signals / wiki_ai_attention_signals, which track fixed article lists over time.

Les titres et les descriptions proviennent des sources de données, en anglais.