Cultural Analytics in R: A Tidy Approach

Table of Contents

1. Frontmatter

2. Chapter 1. Introduction

   Nabeel Siddiqui

   Abstract

   This chapter lays the groundwork for a comprehensive exploration of cultural analytics, an emerging discipline at the intersection of data science and humanities. It begins with a historical overview, identifying the transition from a culture of scarcity to a culture of abundance in the digital age and the challenges and opportunities this presents for humanities research. It advocates for a paradigm shift in humanities scholarship, emphasizing the need for new methodologies to analyze and interpret the vast digital archives now accessible to researchers.

   It introduces the concept of “cultural analytics” as a field that utilizes computational and visualization techniques to uncover insights from massive cultural datasets that transcend the traditional divide between quantitative and qualitative methods. It goes on to explore the specifics of using the R programming language for cultural analysis, highlighting its potential due to its powerful package ecosystem and its ability to handle a variety of data types. The concept of tidy data is presented as a necessary standard for efficient data analysis in cultural analytics, with an emphasis on the “tidyverse” suite of packages that simplifies data manipulation for humanities scholars.

3. Chapter 2. Understanding Base R

   Nabeel Siddiqui

   Abstract

   This chapter provides a comprehensive introduction to R programming. Readers gain a solid understanding of the fundamental concepts of R, ranging from basic mathematical operations and variable assignments to more complex data structures like vectors, data frames, and lists. Through hands-on exploration of a real-world dataset covering 120 years of Olympic athletes, the chapter demonstrates how base R functions can be applied to clean, subset, and analyze data to uncover insights into trends such as health, popularity, nationalism, and gender dynamics. Control structures such as conditional statements and loops are explained, with an emphasis on their practical application in R scripts. Finally, to visually represent data, the chapter covers basic plotting techniques using base R graphics, setting the stage for more advanced visualization tools available in the tidyverse.

4. Chapter 3. Dplyr and TidyR

   Nabeel Siddiqui

   Abstract

   This chapter explores the dynamic realm of cultural analytics, highlighting the transformative impact of the tidy data paradigm on data manipulation and analysis. It serves as a practical guide to the “tidyverse,” with a particular focus on two of its cornerstone packages, “dplyr” and “tidyr.” These tools are instrumental in reshaping complex and messy datasets into an orderly structure suitable for analysis. Through an engaging case study centered on American films from the 1950s to the 2010s, the chapter offers a hands-on demonstration of the power and utility of “tidyverse” functions in real-world scenarios. The case study further explores the merging of disparate data sources through sophisticated joining functions provided by “dplyr.” This extends the analytical framework, allowing the reader to combine film attributes with external datasets such as economic indicators.

5. Chapter 4. Understanding the Grammar of Graphics

   Nabeel Siddiqui

   Abstract

   Delving into the grammar of graphics, this chapter unravels the complexities of data visualization in cultural analytics, emphasizing its interpretative nature. The exposition is grounded in the theoretical underpinnings of “ggplot2,” inspired by Leland Wilkinson’s pioneering framework, illustrating the construction of graphical elements through logical principles that mirror grammatical constructs in language. Key elements of the grammar of graphics are explored—from the manipulation of data and aesthetics to the strategic use of geometric objects, statistics, and themes—equipping researchers with the acumen to translate complex cultural data into persuasive visual stories. The section culminates with the application of “ggplot2” to the Musical Sentiment Dataset (MuSe), using practical examples and exercises to showcase the tool’s capacity for dissecting the emotional impact of music across genres.

6. Chapter 5. Analyzing Text the Tidy Way

   Nabeel Siddiqui

   Abstract

   The rapidly expanding field of cultural analytics has necessitated the development of robust methods for analyzing large volumes of textual data. Central to this endeavor is transforming text, a fundamentally human medium, into quantitative data that machines can process. This chapter details this transformation by applying the tidy text format, simplifying complex text into manageable tokens. Techniques such as sentiment analysis and term frequency-inverse document frequency (TF-IDF) probe deeper into the language, revealing patterns that indicate overall sentiment and word significance respective to distinct contexts. The chapter culminates in exploring topic modeling, specifically latent Dirichlet allocation (LDA), to discern thematic structures within a corpus of text by Mary Shelley, F. Scott Fitzgerald, and Charlotte Perkins Gilman. This analytical process highlights the power of tidy text in unveiling cultural and linguistic patterns and the potential of computational methods in enriching our understanding of cultural phenomena through text.

7. Chapter 6. Understanding Regression

   Nabeel Siddiqui

   Abstract

   The chapter presents an accessible exploration of regression analysis tailored to the cultural research domain. The treatment begins with an overview of linear regression, elucidating the method’s core objectives, assumptions, and its practical utility for inferential statistics and prediction. By leveraging a popular dataset, the chapter illustrates the process of building, interpreting, and validating linear regression models, emphasizing their applicability in drawing robust cultural inferences. The narrative progresses to logistic regression, a technique adept at modeling binary outcomes. Throughout, the chapter balances statistical rigor with practical guidance, equipping readers with the tools to both understand and apply regression analyses to cultural datasets. The chapter details insights into the design intentions of the Pokémon series of games, as inferred through statistical patterns, offering a compelling example of how quantitative methods can yield qualitative understandings in cultural studies.

8. Chapter 7. Tidy Networks

   Nabeel Siddiqui

   Abstract

   This chapter delves into the application of network analysis within the R ecosystem, utilizing the “tidygraph” package to analyze the Early Race Film Database. By leveraging the tidy principles and integrating them with network-specific analysis, it uncovers the intricate web of interactions among films, actors, directors, and production companies. The chapter demonstrates how to import and wrangle network data, calculate centrality measures, visualize connections, and detect communities, revealing the latent structures of cultural networks. It overviews visualization techniques, facilitated by the “ggraph” package, and how they translate complex relational data into insightful diagrams.

9. Chapter 8. Machine Learning with Tidy Models

   Nabeel Siddiqui

   Abstract

   This chapter explores machine learning within the tidyverse in R. It details “tidymodels,” a “meta-package” that streamlines various machine learning workflows, facilitating the use of advanced analytical methods in cultural research. A key focus is placed on decision trees, random forests, and the critical role of hyperparameter tuning. Using a case study from art history, the chapter demonstrates how these methods can uncover biases in the representation of artists across demographics in educational textbooks. Through guided examples, readers gain insights into the iterative nature of model tuning using hyperparameter grids and learn how to assess the accuracy of predictive models.

10. Chapter 9. Conclusion

    Nabeel Siddiqui

    Abstract

    The final chapter of this book emphasizes the boundless nature of cultural analytics, driven by vast datasets and innovative methodologies. It introduces three emerging areas of research—spatial analysis, advanced machine learning, and visual data analysis—that offer new horizons for scholars and practitioners. This section serves as an inspirational springboard, urging readers to extend their knowledge and engage in the dynamic field of cultural analytics beyond the book’s scope.

11. Backmatter