A case study in text mining: Textual analysis of the Territorial Papers.

The Territorial Papers of the United States are a valuable and underused resource containing almost 10,000 documents written between 1789 and 1848 about the formation of new sovereign states from US territory. These communications between the federal government and frontier settlers comprise the actual discourse of the nation's expansion over six decades. Digitizing the Territorial Papers permits the possibility of analyzing the entire corpus globally. Text mining and topic modeling methods give us a lens on the language patterns through which new state governments and the expanding nation were formed. An initial statistical analysis of the textual information provides a visualization of content, helps discern how ideals about governance emerged, and lays the foundation for developing more sophisticated hypotheses and theoretical constructs. [ABSTRACT FROM AUTHOR]