Visual Methods for Digital Research: An Introduction

By Sabine Niederer and Gabriele Colombo

UK and USA, Polity Press, 2024, 201 pages

ISBN-13: 978-1-5095-4254-3

ISBN-13: 978-1-5095-4255-0 (paperback)

One book turning into a bookshelf (virtual)

The opening sentences of Visual Methods for Digital Research: An Introduction (Niederer and Colombo, Polity Press, 2024) describe the ‘digital’ artwork titled Esercizio n.7/Déjà vu (2014)² by artist Caterina Rossato on the cover: ’an assemblage of cut-out landscapes, neatly arranged one behind the other, is held together by two poster clips. The images combined create new and non-existent landscapes that yet seem familiar … The work not only shows a collection of images, reassembled in this case to trigger the sense of déjà vu and play with false memories of landscapes visited. It also is a beautiful example of – to put it very simply – ‘doing things’ with images. (Niederer and Colombo 2024: 1).

Two-page long description of a ‘digital’ artwork (Angela Bulloch’s Horizontal Technicolour, 2002, a throbbing digital light grid of blown-up film footage and solid colours)³ also launches Meredith Hoy’s From Point to Pixel: A Genealogy of Digital Aesthetics (Dartmouth College Press, 2017). Both Rossato’s and Bulloch’s works are what Hoy would call ‘of digital aesthetics’ (Hoy 2017: 11) Even though Rossato doesn’t employ any computational technology, she analogically mimics the experience of digital visual culture (almost like a Google Image Search result), a huge stream of similar, recombinatory and exchangable images, overlapping.

Although both titularly ‘Digital’ in their scope, the Niederer-and-Colombo volume and Hoy volume have vastly different objectives. The former intends to ‘do (research) things’ (Niederer and Colombo 2024: 3) with networked images on the platformized Internet; beyond their production, meaning, circulation and reception. Over six chapters it offers highly imaginative and resourceful data visualization research protocols that are rooted on the specificity of digital visual culture and expansive in terms of various interdisciplinary research goals.

The latter volume by Hoy argues the separation of ‘digital’ and ‘computational’ for digital art from an aesthetic standpoint while trying to trace a pre-(art)history of digital aesthetic in the pre-computer era. (2017: 1)

Yet, the former volume, very surprisingly (because the latter is about digital art itself and the former merely about researching digital images) uses more instances of digital art than the latter – to read digital images closely, distantly, generatively (Generative AI images) as it expands on research protocols, methodological nuances, and theoretical roadblocks. And these artworks are not used as illustrations of methods by Niederer and Colombo, but as inspirations and analytic devices.

The research protocols in the volume – although extremely varied, nuanced and customizable – are essentially advanced data visualizations. Advanced because the visual data which they re-visualize are not the products or results of research but the collected and prepared research material itself. These data visualizations are ‘more than the aesthetic culmination of analysis (Niederer and Colombo, 2019) or…a sort of materialization of media reveals (Allen, 2020) that represent the endpoint of most digital research.’ (Niederer and Colombo 2024: 26) Here they are employed to study a group of images (gathered by specific methodical means): close reading of a small set of images and distant reading of a large set of images. But how the sets are selected and organised are key in these methods. A vivid term for this preparatory act is ‘metapicturing’ (Rogers, 2021) which creates a kind of ‘composite image’ (Colombo, 2018) out of tens, hundreds, thousands and sometimes millions of pre-selected and pre- contextualised images.

For example, on page 16 of the Niederer and Colombo volume, the artwork Reblogs or Context is the new content,⁴ designer- researcher Silvio Lorusso (2015) is discussed by the authors in order to introduce digital image contextualisation and pre-selection methods via source-based, expert-curated, query-based, and image-based techniques. The authors explain Lorusso’s visualization:

The trail created by a single digital image while travelling through a social platform. The project presents a video montage of the same image, … circulating in different online spaces. … accelerates a hectic process of constant modification in meaning and function. While travelling around, images get remixed, downsized, previewed, thumbnailed, or otherwise modified, rendering them “ephemeral” (Hand, 2017), “fugitive and transient” (Lister, 2013). Such instability requires image researchers to develop adaptive methods attuned to the instability and multiplicity of images in flux … Through the visualization of groups of (circulated) images, or “metapicturing”, such poor images can be reunited with their original (Rogers, 2021) and studied as collections.

(Niederer and Colombo, 2019: 17)

What does the ‘data visualization’ method of digital artist Lorusso have in common with the data visualization methods of digital researcher Niederer and Colombo – not in an aesthetic sense (art or non-art), nor ontological sense (what is art? what is method?), neither in processual sense (what is an artist’s method vs what is a researcher’s method) but in a generative epistemological sense – what kind of knowledge they tend to create respectively ?

Hoy does a revelatory comparison in this regard in the Introduction and Conclusion chapters of her volume. She reminds us that data visualization is the prime method of knowledge-making for digital artists, corporate platformised visual culture industry, and academic researchers alike. But the respective methods are worlds apart. For Hoy:

It is precisely the ambiguity (of digital artworks) that permits them to come to rest within the borders of art. By contrast, information visualization…typically achieves success by eradicating ambiguity … Information visualization set itself first to the task of translating numerical data into graphical patterns … would make possible the topographical mapping of the informatic territory in question…. visualization technologies can represent the ‘unrepresentable’; things that could not possibly be seen can be visualized… There is an implicit equation here between visualization and clarity. To visualize a data set within a particular design schema is to make it knowable, sensible, and digestible.

(Hoy 2017, 186)

And later, for example., in our case, the task of data / information visualization is visualizing sets of visuals themselves, adding another dimension.

Yet the two ends of this epistemological binary join each other in circle – like the self-eating serpent Ouroboros – the mythical snake which eats its own tail, at once self-destructing and self-nourishing. The vastness of digital visual data, and the even vaster options to join the visual data dots by myriad methods to make meaning, turn the digital field not very unlike the analog reality itself. The digital artist and the digital researcher get conjoined here. According to Hoy:

While technoscience works toward clarity and disambiguation, art seems to thrive on its irreducible density and ambiguity. And yet, it is in the very impossibility of perfect clarity, articulation, and discretization that the digital becomes an artistic, not merely a technoscientific, communicative form. In seeming to offer the promise of a world that is totally knowable in its utter calculability, the digital ultimately proposes a new set of questions about meaning, a new way of seeing and conceptualizing the world.’

(2017: 209)

Hoy exemplifies,

In Every Playboy Centerfold (2002),⁵ a program written by digital artist, scholar and programmer Jason Salavon compiles a point-by-point digital average of all Playboy centerfold foldouts from January 1988 to December 1997. The seductive poses of these soft-porn icons dissolve into what Salavon calls a “shroudlike,” indistinct form.’ (2017: 195)

A visual data visualization that destroys the overtly visible in the visuals altogether.

Taking Alfred and Aby to Bangladesh (posthumously)

In order to ascertain the ‘Visual vernaculars’ (Niederer 2019; Pearce et al 2020 of various online platforms (search engine, social media and others, leviathan carriers and corpora of circulating images) – how each platform visually formats a specific issue or topic like, say, ‘climate change’ or ‘#beautifulwoman’ – Niederer and Colombo advise the method of stacking. They demonstrate it with the query of the keyword [campus] in Google Image Search. Here translucent versions of all image results are layered on top of one another. It enhances, unlike the averaging out the overt in abovementioned Salavon’s Every Playboy Centrefold, the dominant visual feature of the images. In this case, the resultant image mostly looks like an open lawn. ‘Google Image Search thinks a campus is mostly students sitting together on a lawn’ (Niederer and Colombo 2024, 70).

The key difference between ‘Salavon Averaging’ and ‘Niederer-Colombo Stacking’ seems to be in their utilitarianism of generative epistemology – the usefulness of the respective knowledges they create. The former doesn’t want to deduce any significant feature in its dataset (rather wants to anticlimactically diminish all), contrasting the latter.

While one may simply dismiss that contrast as their generically binaristic behaviour (‘art’ being non-utilitarian and ‘research’ being utilitarian). But historically speaking, their contrast could be argued as not merely generic but generational. Johanna Drucker in several book lengths and bodies of collaborative works argues that the early era of Digital Humanities (now rarely existing as a single academic discipline) in the late 1990s-early 2000s refused the art-science divide. According to Drucker:

For us, the goal of digital work was not to approach the condition of science—quite the contrary, it was to promote humanistic methods. …, aesthetic artifacts were not merely to be objects of study, the material under examination in new digital frameworks, but were to serve as methodological models for creating platforms and procedures in these technologies. (2022: 194)

It is not a coincidence that the examples of digital arts discussed here (and others in their source volumes) are mostly from the 2000s. Two decades later in the 2020s, once again, they are proving relevant to serve as ‘methodological models’.

Here are just two fascinating examples of many such Digital Humanities (DH) projects undertaken in the noughties.

• 1. The ‘Patacritical Demon’ project taken up between 2001 and 2003 at the University of Virginia, USA took its artistic theorisation from French proto-surrealist, dramaturge and pioneer of absurdist philosophy Alfred Jarry (1873–1907). His philosophy of:

  Pataphysics – a science of particular, of exceptionism and imaginary solutions against the probabilistic generalisation – has been argued to be ‘ a response to the dominant models of late nineteenth-century statistics. In traditional quantitative analysis, anomalies were characterized as outliers. They were not seen as indicators of significance, and certainly not as indicators of the greatest significance. (Drucker 2022; 195)

When applied to the core humanities preoccupation of the interpretation of texts in Digital space it offers a trialectic instead of a dialectic model- the significant third along with Thesis and Antithesis is the unique subjective experience of an interpretant, not lapsing into the infinite relativism but holding on to ‘anomalies as points of light shining from within the abyss of the as-yet-unknown, signaling directions for exploration ahead.’ (Drucker 2022, 195)

• 2. German Art Historian and Cultural Theorist Aby Warburg’s (1866–1929) Bilderatlas Mnemosyne (1924–1929)⁶ allows for thinking though images using digital humanities (DH) methods. The Bilderatlas in its final constellation (1929) consisted of sixty-three portable frames (panels) stretched with black cloth onto which black and white photographs (971) of photographs, paintings, sculptures, coins, stamps, popular images and tarot cards, were pinned. It would not be too optimistic to state that Warburg’s Bilderatlas ‘anticipated DH strategies’. (Du Preez 2020, 374).

For Martin Warnke and Lisa Dieckmann:

The project Meta-Image, which is based on the working method of the famous art historian Warburg, … has been under development since 2009 by three universities– the Leuphana University Lüneburg, the University of Cologne and the Humboldt-University of Berlin– and consists of two components: Prometheus, a distributed digital image archive for research and study, and HyperImage, an image linking and annotation tool.

(2016: 109)

Like its ‘Patacritical counterpart, here, out of a seemingly insignificant detail of a picture arises a link in a chain of argument or, by continuing the referencing even further, a nexus in a web of references.’ (Warnke and Dieckmann 2016: 113)

Getting airdropped to an uprising in Bangladesh (digitally)

Can Warburg and Jarry do digital fieldwork with Niederer and Colombo and me in Bangladesh?

Can a Student-People’s Uprising¹ under the banner of ‘Anti-Discrimination’ be a lens for the GSM (Gender and Sexual Minorities) digitality of Bangladesh?

In the bloodied last week of July 2024 marching towards the two-day One-point Movement in Bangladesh⁷ that toppled the incumbent government, a Facebook post was deleted.

Muhin Hashmi (pseud.), a popular Bangladeshi gay satire content creator on Facebook posted in Bengali that the impossible task of uniting all of Bangladesh can only be done around anti-GSM hatred. The unified backlash from all quarters of Bangladeshi gender and political spectrum followed. Hashmi deleted the Facebook post.

At the digital ‘Manosphere’ (Ribeiro et al, 2021) of Bangladesh’s platformised content in Bangla, the GSM idiolect shoehorns itself amongst the ‘circulatory assemblages and representational tropes’ (Mukherjee and Nizaruddin, 2022: 5) of the majoritarian dialect.

Yet, despite being in the lowest rung among South Asian countries as per the LGBT Equality Index, Bangladesh’s production and circulation of platformised GSM-related content (collectively christened Gender Variant Content or GVC) in Bengali – philic, phobic, coded, passing and hate speech – is voluminous and steadily accelerating.

The online crowd politics, activated by young men (often students) – historically both the fuel and the fire in all the recent socio-political-cultural uprisings of Bangladesh – has been paradoxically both progressive, and reactionary. Bangladeshi Gender and Sexual Minorities, trebly bound with religious proscriptions, social stigmatization and constitutional law have had their idiolect actively or passively silenced by the dialect of ‘more urgent’ progressive discourses.

The 2024 Bangladesh quota reform movement that, within weeks, climaxed into the two-days One-point Movement toppling a 16-year-old dictatorship, has created a unique nationwide socio-political-cultural upheaval. The movement reached its goal on the wings of platformised social media content, online (despite repeated Internet blackouts) and hundreds of young (including minors) students shot, maimed, and killed – on ground. Preceded and followed by several major Internet-fuelled nationwide controversies about GSM vis a vis nation-building, this volatile and violent period from June 1 to August 5, 2024 was an interregnum of relative GSM content ‘radio silence’.

Fortified by the methods above, instead of relevant ‘controversy mapping’ online (Munk, Meunier, Venturini 2019) before and after, can we research the ‘radio silence’ triggered by the ‘Anti-Discrimination’ ideologies instead? Can it help us design visual methods for digitally researching the GVC (Gender Variant Contentscape) of Bangladesh?

Taking inspiration from Jarry’s penchant from anomaly and Warburg’s attention to ‘seemingly insignificant detail’ we will approach the aforementioned ‘radio silence’ of GVC in Facebook.

Niederer and Colombo’s volume’s Chapter 3 ‘Networked Images: Platform Image Analysis’ (2024: 48–71) on close reading small sets of images and especially its subchapter ‘Single-Platform Visual Analysis: Studying Multiple Issues on a Single Platform ’(Niederer and Colombo 2024: 52–59) is relevant here. If the research question here is what does the silence of GVC content makers among the huge content surge about the ongoing Student-People Uprising tell us?

Since in this case, the reduced production and circulation GVC content are the ‘sites of meaning’, we have to start our data collection for visual research design against the grain of the platform metrics, yet employing them. The ranking and hashtags of collective images have to be examined from the bottom of the list and not from the top, because we are looking for the covert.

Niederer and Colombo provide some key advice for reading the silent sites of meaning:

Depending on the research question, it may be most fruitful not to predefine the variables and values but, … capture the specific issue language of various actors and sources as they resonate online. … this approach aspires to follow the actors in their own words.

(2024: 53).

The GVC content might have developed new kinds of ‘codes of silence’ in the tumultuous times, they might be missed out if we don’t modify our keywords of search. Studying the content of prominent GVC Influencers closely during this time period might help us here.

The flow of multiple genres of GVC visual content (the ‘pornified’ content for example) whose topic deliberately avoid any mention of the ongoing uprising can be a useful resource for reception studies (platform metrics of like / share / comment for example). In the time of explosive visibilization of all Bangladeshi Facebook content during the uprising, Facebook GVC content might have been finding a losing battle of going ’re-closeted.’ It is up to this research design to test the assumption of this hypothesis.

Competing Interests

The author has no competing interests to declare.

Author Information

Sourav Roy (MPhil, Visual Studies, 2023) is an independent scholar, author, translator and editor currently researching Bengali GVC.