Corpus Analysis: Lexical Dispersion, Semantic Time Series, and Semantic Network Analysis—An R Studio Pipeline

This chapter introduces lexical dispersion analysis, time series analysis, and semantic network analysis. The case study in this chapter uses State of the Union addresses delivered yearly by American Presidents from George Washington to Donald Trump. Lexical dispersion refers to the embedding of words, stems, and n-grams across corpora. Time series analysis is useful when you have a corpus that was produced over the course of hours, days, years, or decades. It enables us to track the prevalence of various words, stems, and n-grams—along with the concepts they express—over time. Time series analysis reveals trends in these speeches across more than two centuries. Semantic network analysis is useful when you have a sufficiently large and diverse corpus to ask which words, stems, and n-grams—along with the concepts they express—tend to co-occur in the same sentence, paragraph, or other piece of text. This can help to shed light on the meaning of terms. These analytical approaches can also be combined to examine how patterns of association change over time.