GDELT daily top world events by coverage

Daily world-events leaderboard from the GDELT Project's keyless 2.1 event update files: the trailing complete UTC day's 200 most-covered individual world events (IsRootEvent=1, GLOBALEVENTID-deduped, ranked by GDELT NumArticles), one row per event. Columns: day-granular date/as_of/fetched_at stamps (any run rebuilding the same window hashes identically), coverage_rank (1 = most covered), the GDELT GLOBALEVENTID, actor1_name/actor2_name as reported by GDELT (unresolved strings, may include spelling variants), the 2-digit CAMEO root event code and its codebook label, QuadClass (1-2 cooperation, 3-4 conflict), AvgTone (-100 very negative to +100 very positive, 2 dp) and Goldstein scale (-10 conflict to +10 cooperation, 2 dp), total_articles (GDELT NumArticles — the number of articles mentioning the event), and the event's country resolved to ISO alpha-3 via the embedded FIPS 10-4 crosswalk (action geo, then actor1 geo, then actor1 code). Primary key: (date, coverage_rank); zero nulls. Method: the connector's shared cleaning pipeline (root-event dedupe, numeric coercion, valid CAMEO roots only, phone-PII screen on free text); days with fewer than 50 clean root events fail loudly rather than publishing a thin leaderboard. Units: counts, tone -100..100, Goldstein -10..10, rank integer. Caveats: actor names are GDELT's raw strings (person/org mentions as parsed, not disambiguated entities — treat as keywords, not identities); coverage is media attention as much as event significance, and GDELT coverage varies by language and over time; NumArticles counts articles mentioning the event, not distinct outlets. Only event metadata and aggregates are stored — no article bodies, no per-row source URLs. GDELT terms require citation of the GDELT Project and a link to https://www.gdeltproject.org/. Sample use: filter quad_class in ('3','4') for the day's most-covered conflict events, or scan coverage_rank 1-20 for the day's headline stories.