"Social processes operate at many temporal scales: decades, years, months, weeks, days, hours, minutes, and even seconds. With the advent of the Internet, social media, and other technologies, long sequences of time-series data are increasingly availableat very fine scales (e.g., an hour of second-by-second recordings produces 3,600 data points; a day of minute-by-minute time-stamped information yields 1,440 data points).In Recurrence-Based Analyses, Sebastian Wallot and Giuseppe Leonardi introduce techniques developed in physics and physiology for characterizing and analyzing patterns in long sequences of temporal data to a broad audience of social scientists. In contrast to time-series regression and other related techniques, recurrence quantificationanalysis (RQA) arises in the context of chaos and nonlinear dynamical systems theory-theory arguably very relevant to social processes. The goal of Recurrence-Based Analyses is to characterize the system's complexity, stability and instability, and conditions under which it transitions from one state to another. The volume opens with an engaging example, a short poem for children entitled "Popcorn" by Helen H. Moore. Although the poem is not a time series per se, it is an ordered sequence of values (letters) that can be seen in this way. Professors Wallot and Leonardi use the repeating sound patterns in this poem to illustrate the concept of recurrence, the construction of a recurrence plot, and a variety of measures that quantify characteristics of thisplot. The poem is short, with lots of rhyme and repetition (pop, pot, hot, top, stop). The recurrence plot, a matrix of the cross comparison of the values of a time series (in this case, letters of the poem), is wonderfully visual. Many measures can be calculated from the recurrence plot, which enables the reader to relate their values to the patterns they can (literally) see in the plot. This first chapter is accessible to all readers. It provides the foundation for the more technical material presentedin subsequent chapters which cover univariate RQA (Chapter 2), techniques for cross-referencing sequences (Chapters 3-5), and extensions to analyze more than two series at once (Chapter 6)"-- Provided by publisher.
Sebastian Wallot and Giuseppe Leonardi introduce techniques developed in physics and physiology for characterizing and analyzing patterns in time series data to a broad audience of social scientists. In contrast to time-series regression and related techniques, recurrence quantification analysis (RQA) has its background in chaos and nonlinear dynamical systems—theory arguably very relevant to social processes. The goal of Recurrence-Based Analyses is to introduce readers to these techniques that can characterize a system’s complexity, stability and instability, and conditions under which it transitions from one state to another. The authors illustrate concepts and techniques with relevant social science examples at different temporal scales: biweekly polling data on federal elections in Germany; daily values of three stock market indices; daily cases of SarsCov-19 in four countries during the pandemic; and second-by-second vocalizations of mothers and infants interacting recorded by motion cameras. This introduction to RQA serves as a useful supplement to undergraduate and graduate courses in computational social science, and also by researchers who seek new tools to address social scientific questions in new ways.
