Significance
15 measured / 15 eligible · 0 missing · 0 not applicable
score (0 to 1) · Event-level distribution, not a Situation risk score.
across all 15 events
Russia carried out drone strikes across Kyiv region on 2026-09-20, killing four people, in “Russian forces strike Ukraine’s logistics hubs, warehouses and ports overnight.” Russian Armed Forces struck an Eco Market warehouse in Zhytomyr region on 2026-09-20, identifying civilians as the target and reporting no fatalities. Russian drones also struck railway infrastructure in Vinnytsia region on 2026-09-20. On 2026-09-21, Russian Armed Forces struck critical infrastructure in Poltava, a Ukrainian military logistics center in Kyiv Oblast, and an enterprise in Kropyvnytskyi district, where one person was injured. The same day, “Russian strikes hit apartment buildings and shopping center in Zaporizhzhia” and “Russia strikes educational facilities and shopping centers in Zaporizhzhia” described civilian-area damage, while “Attack damages Ukrainian rail infrastructure, delaying 50 train services” described disruption to rail services.
Footnotes[7][2][8][11][10][12][9][1][6][3]The coded Events and linked Stories describe Russian air-and-drone strikes affecting civilians, infrastructure, and Ukrainian military or state-force targets across Ukraine. Same-day links connect the logistics Story with the rail-disruption and Zaporizhzhia civilian-damage Stories. The six coded Events identify Russia or Russian Armed Forces as the source actor. The material includes reported deaths, an injury, rail delays, and strikes on infrastructure and logistics sites; the same-day Stories are not all separate coded Events. The targeting classifications are not uniform: the Zhytomyr warehouse Event marks civilian targeting true, while the Kropyvnytskyi enterprise Event names civilians as the target but marks civilian targeting false.
Footnotes[17][20][22][23][24][8][9]Explanation originally captured 2026-09-22 · Evidence through 2026-09-22 UTC · Source edition sed_baadb2ffb34d39f9416a0a8221152c93
A 3-day Situation represented by 15 coded Events carried across 29 member Stories.
The coded activity is led by AIR/DRONE STRIKE (9) and SHELLING/ARTILLERY/MISSILE ATTACK (2).
Donald Trump is prominent in the retained actor and relationship evidence.
115 canonical Entities are linked to this edition.
Metric coverage. 15 coded Events · 3 CAMEO+ · 12 conflict. The four dimensions describe the CAMEO+ subset; conflict Events do not carry these scores.
Dot = mean · shaded span = observed range · position uses each metric’s defined scale.
How the coder scored the full retained Event set
15 coded events — 3 CAMEO+ (1 political) · 12 conflict
Quad class over those 13 events · Material Conflict 11 · Verbal Conflict 1 · Verbal Cooperation 1. 2 sit outside POLITICAL, where Goldstein and quad class are undefined. The four CAMEO+ scores are over the 3 CAMEO+ events only; the 12 conflict events have no such column.
15 measured / 15 eligible · 0 missing · 0 not applicable
score (0 to 1) · Event-level distribution, not a Situation risk score.
across all 15 events
2 measured / 3 eligible · 1 missing · 12 not applicable
score (0 to 10) · Event-level distribution, not a Situation risk score.
across the 3 of 15 events coded in the CAMEO+ family; the conflict family has no such column
2 measured / 3 eligible · 1 missing · 12 not applicable
score (0 to 1) · Event-level distribution, not a Situation risk score.
across the 3 of 15 events coded in the CAMEO+ family; the conflict family has no such column
2 measured / 3 eligible · 1 missing · 12 not applicable
score (0 to 1) · Event-level distribution, not a Situation risk score.
across the 3 of 15 events coded in the CAMEO+ family; the conflict family has no such column
2 measured / 3 eligible · 1 missing · 12 not applicable
score (0 to 1) · Event-level distribution, not a Situation risk score.
across the 3 of 15 events coded in the CAMEO+ family; the conflict family has no such column
13 measured / 13 eligible · 0 missing · 2 not applicable
score (-10 to 10) · Event-level distribution, not a Situation risk score.
across the 13 of 15 events Goldstein is defined on — CAMEO+ POLITICAL and the conflict family; the other nine CAMEO+ domains have no Goldstein semantics
12 measured / 12 eligible · 0 missing · 3 not applicable
reported people · Reports may overlap; these are Event-level reports, not an additive unique human toll.
across the 12 of 15 events that report a toll — the conflict family only, and never 0 on a CAMEO+ event
15 measured / 15 eligible · 0 missing · 0 not applicable
score (0 to 1) · Event-level distribution, not a Situation risk score.
across all 15 events
This retained edition, independent of graph and page limits. Different dimensions, not a blended risk score; reported tolls may overlap.
2026-09-22from the recordslatest
Chernihivoblenergo reports power outage affecting 100,000 subscribers
23 Stories and 7 coded Events (+17 / +2 against the day before). Widest coverage: Russian strikes hit industrial sites in Dnipro, Kryvyi Rih and Pavlohrad (68 articles).
What kind of event · 6 coded types
over 15 coded events on this edition
Countries · where it happened
Countries · who is acting
Selected from the complete retained edition
Russia or Russian Armed Forces are identified as the source actor in the Kyiv-region, Zhytomyr, Vinnytsia, Poltava, Kyiv-Oblast and Kropyvnytskyi Events.
The Events report four deaths in Kyiv region, one injury in Kropyvnytskyi district and a civilian-targeted warehouse strike in Zhytomyr; the Zaporizhzhia Stories describe apartment buildings, shopping centers and educational facilities.
The logistics Story covers warehouses, ports and logistics hubs; coded Events separately identify railway infrastructure in Vinnytsia, critical infrastructure in Poltava and a military logistics center in Kyiv Oblast.
The rail Story reports delays affecting 50 train services, while the power-outage Story attributes outages to a Russian drone attack and bad weather.
Recorded as actor in 1 coded Event, including Trump calls for Ukraine war to end.
Recorded as actor in 1 coded Event, including Trump calls for Ukraine war to end.
Footnotes[?]

On 2026-09-20, Russian Armed Forces struck an Eco Market warehouse in Zhytomyr region, with civilians identified as the target and no fatalities reported.

On 2026-09-20, Russian drones struck railway infrastructure in Vinnytsia region, with Ukrainian state forces identified as the target.

On 2026-09-21, Russian Armed Forces struck critical infrastructure in Poltava; no fatalities or injuries were reported.

On 2026-09-21, Russian drones struck a Ukrainian military logistics center in Kyiv Oblast, with the Ukrainian Armed Forces identified as the target.
Evidenced relationships involving significant participants beyond the central cast.
The logistics Story and the rail Story are connected by a same-day link; the rail Story reports damage and delays affecting 50 train services.
The logistics Story and the Zaporizhzhia Story are connected by a same-day link, with the Zaporizhzhia Story describing apartment buildings and a shopping center.
The power-outage Story is linked to the logistics Story as same-day infrastructure reporting; it attributes outages to a Russian drone attack and bad weather and has no linked coded Event.
The Kyiv-Oblast military logistics-center Event lists the logistics Story among its linked Stories.
The Vinnytsia railway-infrastructure Event lists the rail Story among its linked Stories.
This edition freezes the complete served membership and analysis available through its UTC cutoff. Later repairs or identity changes do not rewrite what this edition showed.
Stories are reporting clusters. Events are coded occurrences within that reporting. The cast contains 115 canonical people and organizations linked to the retained evidence.
Story-to-Event edges identify supporting reporting. Entity-to-Event edges identify recorded actor or target participation. Mentions and analytical roles stay visibly distinct from recorded participation.
Core, precursor, and continuation labels organize the reading. They do not by themselves establish causation. Collapsed graph groups and paginated evidence change presentation only; counts use the complete retained edition.