polecat_events — Political events between actors (POLECAT, 2018-2024)

Description

A temporal network of political events between socio-political actors worldwide (governments, politicians, parties, the military, the police, armed groups, journalists, citizens, ...), from January 2018 to June 2024, from the POLitical Event Classification, Attributes, and Types (POLECAT) dataset, the successor of the Integrated Crisis Early Warning System (ICEWS) event data. POLECAT codes events automatically, with a machine-learning coder (NGEC) based on the PLOVER ontology of event types, from news stories in seven languages (English, Russian, Spanish, French, Portuguese, Arabic and Chinese) from over a thousand sources. Each directed edge from A to B is an event in which actor A acts upon actor B (e.g. A accuses, threatens, aids or assaults B), with its date, type and other attributes.

Each event lists its actors and recipients (possibly several of each). An actor with a name in the data (the name resolved by the coder, usually that of a person or organization, e.g. Narendra Modi, NATO) is a node identified by that name. Actors without a name, which are described only by their country and their sector (e.g. the Ukrainian military, the Indian police), are nodes identified by the country and the sector, e.g. Ukraine MIL or India COP; actors with neither name, country nor sector are left out. An event with several actors or recipients becomes one edge for each pair of actor and recipient. Events without any identifiable actor or recipient (about a third of the 2.2 million events) are not included. Of the 9 event identifiers that appear twice in the data, only the first occurrence is kept. The data are the yearly files of version 62 of the "POLECAT Weekly Data" dataset (the file of 2024 ends in June 2024). The raw names and titles of the actors as they appear in the stories, the people, organizations and locations mentioned in the stories, and the finer geographic information (city, district, province, GeoNames identifiers) are not included.

The network has the following properties:

(Node) name: the name of the actor, or its country and sector for unnamed actors.
(Node) named: whether the actor has a name in the data (otherwise it is defined by its country and sector).
(Node) country: the country of the actor (the most frequent one in its events; e.g. Intergovernmental Organization for international organizations; None if not known).
(Node) cow_code: the Correlates of War code of the actor's country (the most frequent one; -1 if not known).
(Node) sector: the primary sector of the actor (the most frequent one), as a PLOVER sector code (e.g. GOV government, MIL military, COP police, JUD judiciary, LEG legislature, PTY political party, REB rebels, CVL civilians, JRN journalists, BUS business; None if not known).
(Node) wikipedia_id: the title of the Wikipedia page of the actor, as resolved by the coder (the most frequent one; None if not known).
(Edge) event_id: the identifier of the event.
(Edge) time: the date of the event, in Unix time.
(Edge) event_type: the PLOVER type of the event: AGREE, CONSULT, SUPPORT, CONCEDE, COOPERATE, AID, RETREAT, REQUEST, ACCUSE, REJECT, THREATEN, PROTEST, SANCTION, MOBILIZE, COERCE or ASSAULT.
(Edge) event_mode: the PLOVER mode of the event, a subtype (e.g. arrest, demo, ceasefire; empty if none).
(Edge) intensity: the intensity of the event, from -10 (most hostile) to 10 (most cooperative).
(Edge) quad_code: VERBAL COOPERATION, MATERIAL COOPERATION, VERBAL CONFLICT or MATERIAL CONFLICT.
(Edge) contexts: the contexts of the event (e.g. military, election, health), separated by | (None if none).
(Edge) placename: the name of the location of the event (None if not known).
(Edge) location_country: the ISO 3166-1 alpha-3 code of the country of the location (None if not known).
(Edge) latitude, longitude: the coordinates of the location (NaN if not known).
(Edge) source: the news outlet of the story.
(Edge) language: the language of the story.

Tags

Social Political Conflict Unweighted Multigraph Timestamps Metadata

Added

2026-10-27

Citation

G. I. Scarborough, B. E. Bagozzi, A. Beger, J. Berrie, A. Halterman, P. A. Schrodt, J. Spivey, "POLECAT Weekly Data." Harvard Dataverse, V62 (2024) 10.7910/DVN/AJGVIT Sci-Hub
A. Halterman, P. A. Schrodt, A. Beger, B. E. Bagozzi, G. I. Scarborough, "Creating Custom Event Data Without Dictionaries: A Bag-of-Tricks." arXiv:2304.01331 (2023) 10.48550/arXiv.2304.01331 Sci-Hub

Upstream URL OK

https://doi.org/10.7910/DVN/AJGVIT
https://doi.org/10.7910/DVN/LMFPIP

Upstream license

CC0 1.0, http://creativecommons.org/publicdomain/zero/1.0

Networks

Tip: hover over a column header for a description.

Column legend

$\left<k\right>$: Mean degree
$\sigma_k$: Standard deviation of the degrees
$\lambda_h$: Largest eigenvalue of the Hashimoto (non-backtracking) matrix
$\tau$: Random walk mixing time
$r$: Degree assortativity
$c$: Global clustering coefficient
$\oslash$: (Pseudo-) diameter
$S$: Relative size of the largest component
NPs: Node properties
EPs: Edge properties

Name	Nodes	Edges	$\left<k\right>$	$\sigma_k$	$\lambda_h$	$\tau$	$r$	$c$	$\oslash$	$S$	Kind	Mode	NPs	EPs	gt	GraphML	GML	csv
polecat_events	91,314	1,996,186	21.86	1081.24	318.37	25.51	-0.07	0.35	9	0.97	Directed	Unipartite	name named country cow_code sector wikipedia_id	event_id time event_type event_mode intensity quad_code contexts placename location_country latitude longitude source language	35.6 MiB	43.7 MiB	41.6 MiB	49.5 MiB

Ridiculograms^*

^* These are automatically generated force-directed visualizations, and can be quite meaningless for networks both big and small. They should not be taken seriously as sources of scientific insight. See here for a discussion.