Phoenix Dataset

The Phoenix dataset is a new, near real-time event dataset created using the next-generation event data coding software, PETRARCH. The data is generated using news content scraped from over 400 sources. This scraped content is run through a processing pipeline that produces coded event data as a final output. Our current settings produce roughly 3,000 coded events per day. These coded events are in the standard who-did-what-to-whom format typically associated with event data. Each event is coded along on multiple dimensions, specifically source and target actors and event type. These dimensions are described in greater detail below.

Event Coding

Each coded event has three main parts: the source actor, the target actor, and the event type. Event types are coded according to the Conflict and Mediation Event Observations (CAMEO) framework. CAMEO consists of twenty top-level codes, and over 200 event codes. These codes can be further aggregated to the commonly used "QuadClass" categories, which are desribed in further detail in the next section.

The actor codes used consist of three character strings grouped together. For example, a sentence mentioning Syrian rebels would have an actor coding of SYRREB. These three character codings can be strung together to create actor codes of arbitrary length. Within the Phoenix dataset, each actor is divided into three types: ActorRole, ActorEntity, and ActorAttributes. The ActorRole is one of either a CAMEO country code, or IGO, NGO, IMG, orMNC. The ActorEntity is one of GOV, MIL, REB, OPP, PTY, COP, JUD, SPY, MED, EDU, BUS, CRM, or CVL. ActorAttributes encompasses any other actor code not covered by the previous two categories. The dashboards shown on this page make use of ActorRole variable for subsetting and aggregation.


The QuadClass variable presents a higher level of aggregation for the CAMEO categories. Unlike previous iterations of the QuadClass conversions, our current implementation adds a "neutral" category to handle events that do not have a strong leaning in a coooperative or conflictual direction. Our current conversion is shown in the following table.

Root CAMEO Description Quad Class Quad Class Description
01 Make Public Statement 0 Neutral
02 Appeal 0 Neutral
03 Express Intent to Cooperate 1 Verbal Cooperation
04 Consult 1 Verbal Cooperation
05 Engage in Diplomatic Cooperation 1 Verbal Cooperation
06 Engage in Material Cooperation 2 Material Cooperation
07 Provide Aid 2 Material Cooperation
08 Yield 2 Material Cooperation
09 Investigate 3 Verbal Conflict
10 Demand 3 Verbal Conflict
11 Disapprove 3 Verbal Conflict
12 Reject 3 Verbal Conflict
13 Threaten 3 Verbal Conflict
14 Protest 4 Material Conflict
15 Exhibit Force Posture 4 Material Conflict
16 Reduce Relations 3 Verbal Conflict
17 Coerce 4 Material Conflict
18 Assualt 4 Material Conflict
19 Fight 4 Material Conflict
20 Use Unconventional Mass Violence 4 Material Conflict