Book Review: 'Visual Methods for Digital Research: An Introduction' by Niederer and Colombo, reviewed by Gerry Derksen

Visual Methods for Digital Research: An Introduction
By Sabine Niederer and Gabriele Colombo
July 2024
256 pages
216 x 138 mm / 9 x 5 in
Polity Press

Reviewed by Gerry Derksen

In an era dominated by visual communication, “Visual Methods for Digital Research: An Introduction” by Niederer and Colombo emerges as a timely and comprehensive guide to navigating the complexities of image-based research. This book offers multiple approaches to understanding, analyzing, and utilizing digitized and generated images in academic and practical research contexts, making it an invaluable resource for scholars, data scientists, and visual communication professionals.

The authors begin by introducing readers to various image types, including non-human, soft, glitch, and networked images. This taxonomy sets the stage for a deeper exploration of how different image categories function in research and communication. Particularly intriguing is the discussion of networked images, which considers the evolution and spread of memes, highlighting the dynamic nature of visual communication in digital contexts.

One of the book’s strengths lies in its presentation of five strategies for studying groups of images: source-based, expert-curated, query-based, snowballing keywords, and image-based collection building. These methodologies provide researchers with a robust toolkit for approaching image analysis from multiple angles. The authors’ explanation in Chapters 2, “Distance Images: Reading Large Collections,” and Chapter 3, “Networked Images: Platform Image Analysis,” describe how these strategies can be applied in various contexts, such as distant and close reading, demonstrating the book’s practical value.

The concept of “distant reading” of images is particularly prescient given the increase in user-generated visual content and collections of large data sets. By discussing how extensive collections of images can be analyzed based on formal similarities, what Lev Manovich calls ‘cultural analytics’ (2012) is applied initially to art and magazine covers and, more recently, used on social media images. The authors introduce readers to powerful organizational techniques using computer vision, or structuring data around timestamps, or digital features such that a hierarchy is formed based on top-level searches to links and sub-linked items. Their example of using clustered data to study images related to the Paris Climate Agreement and subsequent US withdrawal illustrates how these methods can yield insights into complex social issues. What is not mentioned, however, are cautionary tales of user-generated images or tagged associated data that include biases, which the authors only discuss in the final chapter on AI-generated content. Here, Niederer and Colombo highlight the relationship of clustered images, examining the associations between clusters using sentiment from keyword descriptions, for example. The edges of adjacent clusters form new contexts defined by publics who assigned the keywords on a given platform.

Equally important is the book’s treatment of “close reading” of images. The authors advocate for a nuanced approach that considers not just the content of individual images but also their context within metadata, such as ranking, hash terms, search hierarchies, and keyword use, for grouping and overlaying images across smaller collections. This multi-dimensional analysis framework encourages researchers to look beyond surface-level interpretations and consider the broader sociocultural implications of the collection.

Chapter 4, “Critical Images: Exposing Inequalities with Visual Research,” stands out for its engagement with pressing social issues. By examining how images can reveal or perpetuate inequalities, the authors demonstrate how visual research used for social critique can uncover messaging that borders on stereotyping. The discussion of pregnancy vs. unwanted pregnancy imagery and the analysis of Getty’s “Lean In” collection offer compelling examples of how visual representation intersects with social justice issues, even when conscientious efforts are taken to lessen hegemonic influences.

The book takes a significant turn in Chapter 5, “Participatory Images: Talking Back to Maps,” exploring the potential for more inclusive and reflexive research methodologies. By introducing concepts like “matters of care” used by data feminists, the authors challenge traditional power dynamics in research and advocate for approaches that include participants’ perspectives. This section raises crucial questions about authorship and bias in data collection, explicitly encouraging readers to examine their methodologies while advocating for participatory research designs critically. Niederer and Colombo challenge earlier research methods’ inequity as a bias but call for a shift in perspective from power and dominance to emotion and embodiment. This shift localizes the relationship between participant and data visualizations rather than attempting to stress ‘one single loud, or technical, or magical’ voice (D’Ignazio and Klein, 2020). Using current digital visualization techniques, the authors illustrate how images can be seen both as networked and as individual contributions within a participatory research framework.

One of the more topical chapters of the book is Chapter 6, “Machine Images: Generative Visual AI for Research,” which discusses methods of image generation and analysis methods. The distinction between prompt engineering and prompt design offers a nuanced understanding of how to leverage AI in visual research. The discussion of negative prompts, generic prompts, and abstract prompting provides valuable insights into the potential and limitations of AI in image generation and model insights as a focus of study.

While the book covers an impressive range of topics, some readers might find certain sections overly brief. For instance, the two paragraphs on Forensic Architecture leave the reader wanting more about this intriguing approach to crisis analysis. Additionally, while the book raises important ethical considerations, case studies describing their outcomes, from ethical implications of AI-generated imagery to participatory research methods, would have been welcome. Understandably, methods are used differently in studies; however, these examples provide hints as to when and for what purpose they can be applied.

Despite these minor limitations, Visual Methods for Digital Research succeeds in providing a comprehensive and thought-provoking overview of contemporary visual research methodologies. Its interdisciplinary approach, combining elements of data science, social science, and media studies, makes it relevant to a broad audience. The authors’ emphasis on critical thinking and ethical considerations ensures that readers are equipped not just with methodological frameworks but also with the ability to reflect on the broader implications of their research.

Niederer and Colombo have produced a valuable resource that will undoubtedly shape the process of visual research. By encouraging researchers to consider various approaches, engage with emerging technologies, and remain mindful of the social impact of their work, Visual Methods for Digital Research starts a valuable conversation for scholars in the use of these methods and the perspective position from where they originate. It is essential reading for anyone seeking to understand and harness the power of images in academic and applied research contexts.