Crowdsourcing the Historical Record: Creating Linked Open Data for Chinese History at Scale

Abstract

An important part of the historical record of premodern China is recorded in historical works such as the standard dynastic histories. These works are a key source of knowledge about many aspects of premodern Chinese civilization, including persons, events, bureaucratic structures, literature, geography and astronomical observations. While many such sources have been digitized, typically these digitized texts encode only literal textual content and do not attempt to model the semantic content of the text. Similarly, while some of the historical data contained in some of these sources has been entered into specialist scholarly databases, an even greater proportion of the information does not yet exist in any machine-readable form. Producing such a machine-readable dataset of these materials requires the effort of many individuals working together due to the large scale of the task. This article introduces a crowdsourced approach in which annotation and knowledge base construction are carried out in parallel, with a knowledge base continually expanded through multi-user contributions to textual annotation immediately and automatically feeding back to provide improved assistance with subsequent annotation. The resulting knowledge base is dynamically exposed through Linked Open Data interfaces, creating a continually expanding machine-readable dataset covering around 3,000 years of recorded history.