How can we study the social moving image on TikTok? To which extent can we repurpose TikTok metadata for the analysis of networked video cultures? Which new forms of memeification and templatability can we identify? While visual social media analysis can draw on a series of established methodological protocols for working with static images (Rose 2023), researching online video platforms raises new empirical questions.
During this six-hour workshop, we invite participants to explore and collaboratively develop TikTok methodologies by “metadating” and “metapicturing” collections of TikTok videos through their different networked characteristics—original and listed sounds, hashtags, effects, duets, emojis, and stickers. If metadata, according to Lev Manovich, is what allows computers to “connect data with other data” (2002), “metapicturing” (Rogers 2021) can be seen as an analytical technique for inventive, ethical, and contextual data remix. Presenting a set of ‘online-grounded’ possibilities for arranging content into “pictures of pictures” (Mitchell 1995), this hybrid approach allows for making sense of TikTok videos in their multimodality.
The workshop invites media and visual studies scholars to critically engage in practical tutorials and methodological discussions on the quali-quantitative analysis of TikTok video memes. Retaining sensitivity to the contextual nuance of memetic (sub)cultures, methodological affordances of digital research tools tested in a series of Digital Methods Initiative projects (Bainotti et al. 2022; Geboers et al. 2022; Pilipets et al. 2023) will be introduced in three parts:
00:00 - 00:15 Introduction
*00:15 - 01:45 Listed Sounds: Metapicturing Audiovisual Content with Lucia Bainotti and Stijn Peeters/UvA*
In this tutorial, we present how to analyze TikTok audio-visual content and detect the presence of memetic templates by “following the sound” and arranging audio-visual content in a specific type of metapicture—“video stripes”—with the 4CAT Capture and Analysis Toolkit (Peeters & Hagen 2022). Video stripes are horizontal collages of selected sequential frames extracted from TikTok videos. By displaying dynamic video content statically, this technique is developed to highlight the visual patterns and gestural forms assumed by TikTok audio memes.
The entry point for the analysis is a specific platform affordance: the listed sounds indexed by TikTok, which users utilize as templates for their content (Abidin & Kaye 2021). Participants will learn how to repurpose this affordance and "follow" these sounds (cf. Rogers 2019) to investigate associated visual vernaculars and memetic practices of templatability, hijacking, and subversion.
01:45-02:00 Coffee Break
*02:00-03:30 Disguised speech templates: Video stacking and ethical fabrication with Elena Pilipets and Jason Chao/University of Siegen*
In this tutorial participants will learn how to analyze a TikTok video collection with particular attention to the interplay of video effects, duets, and embodied memetic production through speech. We will first learn how to extract static frames from videos using Video Frame Extractor (Chao 2022). With the aid of Speech-To-Text Converter (Chao 2022), we will then demonstrate how to recognize disguised speech templates in videos that were published with the polyphonic “original sound”. Finally, we will visualize a subset of selected videos by repurposing techniques of analytical display known as image stack and montage.
A stack here is not only “a sort of computationally generated moodboard” (Colombo 2018) but arguably also a method of “ethical fabrication” (Markham 2012) that allows for bricolage-style montage and transfiguration of the original images.
03:30-04:00 Lunch Break
*04:00-05:30 (Con)Textual analysis of co-hashtags, emojis and video stickers with Marloes Geboers/UvA*
Within soundscapes, we shift emphasis to the textual signifiers of meaning as provided by combinations of co-hashtags and stickers. The latter allows creators to embed texts and emojis within videos. They can align with the expressive meaning of the visual content and particular sounds, but they can also subvert, or infuse videos with ambiguity. We will discuss a series of methodological designs for contextualizing video content through the analysis of linkages between visual and textual elements.
Similar to part I and II, we will work with image montage outputs. Only this time, we sample videos using the textual data dimensions provided by co-hashtags and stickers. We follow research designs that attune to various research aims, ranging from detecting tactical activities to mapping discursive temporal dynamics. In this way, we repurpose discursive patterns for the interpretational work involved in assessing the meta-picture.
05:30-06:00 Wrap-up: What else can we do to study TikTok video memes?
*Aims and requirements*
Up to thirty participants will work together with workshop facilitators in an interdisciplinary setting, combining qualitative interpretative protocols of close- and cross-reading with data-intensive methods of visual design and storytelling. All participants are required to bring a laptop. TikTok datasets and step-by-step walkthrough documents will be provided as a basis for exploration and further methodological development. No technical knowledge is required.
