Historians and political scientists commonly read large quantities of text to construct an accurate picture of significant events. We have developed a method to help historians identify possible events from the texts of historical communication. This is a tool to browse the results of our method on historical US State Department cables from the 1970s.
This data was obtained from the History Lab at Columbia University.
Please see the Capsule project website for more details, source code, and relevant publications.
We recommend starting your exploration at one of our three overview pages:
Cable: Diplomatic text messages sent to and from the US State Department and its embassies. The word cable comes from a time when physical cables were used to relay the messages. See Wikipedia's page on diplomatic cables.
Topic: An automatically uncovered theme in a collection of documents, or a distribution over words. See David Blei's page on topic modeling for more general resources on topic modeling and our project website for details on the model presented here.
Entity: A sender or recipient of a cable. The US State Department, its embassies, other US government organizations, and sometimes even individuals are entities. Our model describes entities in terms of their typical concerns, or a distribution over general topics and an entity-specific topic.
Event: Inherently an ambiguous conceptual contruct, we define events to have both a start date (or time) and a topic associated with the time interval.