Data Analytics for Discourse Analysis with Python The Case of Therapy Talk Routledge Studies in Linguistics Series

This concise volume, using examples of psychotherapy talk, showcases the potential applications of data analytics for advancing discourse research and other related disciplines.

The book provides a brief primer on data analytics, defined as the science of analyzing raw data to reveal new insights and support decision making. Currently underutilized in discourse research, Tay draws on the case of psychotherapy talk, in which clients? concerns are worked through via verbal interaction with therapists, to demonstrate how data analytics can address both practical and theoretical concerns. Each chapter follows a consistent structure, offering a streamlined walkthrough of a key technique, an example case study, and annotated Python code. The volume shows how techniques such as simulations, classification, clustering, and time series analysis can address such issues as incomplete data transcripts, therapist?client (a)synchrony, and client prognosis, offering inspiration for research, training, and practitioner self-reflection in psychotherapy and other discourse contexts.

This volume is a valuable resource for discourse and linguistics researchers, particularly for those interested in complementary approaches to qualitative methods, as well as active practitioners.

Introduction

Defining data analytics

Data analytics for discourse analysis

The case of psychotherapy talk

Outline of the book

Quantifying language and implementing data analytics

Quantification of language: word embedding

Quantification of language: LIWC scores

Introduction to Python and basic operations

Chapter 2 Monte Carlo simulations

Introduction to MCS: bombs, birthdays, and casinos

The birthday problem

Spinning the casino roulette

Case study: Simulating missing or incomplete transcripts

Step 1: Data and LIWC scoring

Step 2: Simulation runs with a train-test approach

Step 3: Analysis and validation of aggregated outcomes

Python code used in this chapter

Chapter 3 Cluster analysis

Introduction to cluster analysis: creating groups for objects

Agglomerative hierarchical clustering (AHC)

k-means clustering

Case study: Measuring linguistic (a)synchrony between therapists and clients

Step 1: Data and LIWC scoring

Step 2: k-means clustering and model validation

Step 3: Qualitative analysis in context

Python code used in this chapter

Chapter 4 Classification

Introduction to classification: predicting groups from objects

Case study: Predicting therapy types from therapist-client language

Step 1: Data and LIWC scoring

Step 2: k-NN and model validation

Python code used in this chapter

Chapter 5 Time series analysis

Introduction to time series analysis: squeezing juice from sugarcane

Structure and components of time series data

Time series models as structural signatures

Case study: Modeling and forecasting psychotherapy language across sessions

Step 1: Inspect series

Step 2: Compute (P)ACF

Step 3: Identify candidate models

Step 4: Fit model and estimate parameters

Step 5: Evaluate predictive accuracy, model fit, and residual diagnostics

Step 6: Interpret models in context

Python code used in this chapter

Conclusion

Data analytics as a rifle and a spade

Applications in other discourse contexts

Combining data analytic techniques in a project

Final words: invigorate, collaborate, and empower

Dennis Tay is Professor at the Department of English and Communication, the Hong Kong Polytechnic University. He is Co-Editor-in-Chief of Metaphor and the Social World, Associate Editor of Metaphor and Symbol, Academic Editor of PLOS One, and Review Editor of Cognitive Linguistic Studies. His recent Routledge publication is Time Series Analysis of Discourse: Method and Case Studies (2020).