Facilitators: Matt Canute (Simon Fraser University), Mel Racho (York/Toronto Met U), Rory Sharp (York/Toronto Met U), Sasha Akhavi (York University), Ganaele Langlois (York University), Anthony Burton (Simon Fraser University), Cairg Fahner (York University). Matt Canute and Ganaele Langlois are co-principal investigators for the Digital Research Methods stream of the Data Fluencies Project (Mellon-funded), hosted at the Digital Democracies Institute (Simon Fraser University). All other facilitators are research assistants for the Digital Fluencies project. For this workshop, all software and data collected will be hosted on the CEDAR super-computer at Simon Fraser University.
Participants: 25 max
Length: 5 hours.
Though incredibly common, the circulation of images on social media platforms is difficult to study. The reasons for this are technical as well as conceptual. There is an urgent need to further explore the visceral and affective impacts of the endless flow of images users see online (Schlag, 2018). The many visual genres online, from photography to meme (McSwiney et al., 2021), from still to gif (Martinez, 2019), add complexity to researchers’ efforts to understanding role played by images in problematic informational environments, such as in mis/disinformation. And the growing capacity to automatically create visual content through AI only exacerbates this theoretical and methodological conundrum (Crawford and Paglen, 2021). However, image collection requires large storage capacity and specific scraping scripts.
This workshop will showcase open access software tools and methods to assist researchers in exploring the networked, cultural, and affective impacts of images on social media platforms (Dvorak and Parikka, 2021). We start from the premise that overall, images online articulate together new technical affordances with social and cultural impacts: they are mobilized to generate economies of attention and distraction, to cultivate, manage and in turn disorganize affects, defining horizons of existence, shaping interpretations of the worlds (Juris, 2008). This workshop will present a series of quantitative, qualitative, and exploratory software-assisted approaches to examine the role played by images on social media. These software-assisted methods can be used separately or in conjunction with each other. The objective of this workshop is to showcase multi-methodological strategies, from large scale scraping and visualization to the curation of smaller image samples, from formal visual analysis to participatory methods for exploring the relationships between images and affect. We will be teaching workshop participants to use the following tools, going through the steps of collecting images from social media platforms and exploring them using top-down and bottom-up approaches (Davila, 2019):
“Zeeschuimer ”. We have adapted this open-source web browser plugin created by the Digital Methods Initiative that allows for image and metadata scraping of Instagram and TikTok feeds. Participants will use this tool in conjunction with the research persona method (Bounegru et al., 2022), where fictitious user profiles are curated to elicit algorithmic recommendation processes.
“Image Flow” is an open-source tool with the ability to extract and collect images from social media websites such as reddit, twitter, 4chan, and Facebook public groups using varying degrees of search terms. This tool enables more thematic collecting of images.
Exploratory Image visualization: A key capacity of Image Flow is to enable the visualization of clusters of images (using an adapted version of PixPlot). Using images collected from both Zeeschuimer and Image Flow, we will visualize images clustered by semantic image similarity, comparing the spread of images. This tool allows for a view from the top, which we will contrast with a view from the bottom with the last tool below.
“Image and Affect” is an open-source and collaborative browser plugin tool that enables users to record their affective states while encountering images on their social media feeds. This exploratory approach responds to the immense importance of visual information in relation to affect production and circulation that is otherwise overlooked by strictly textual approaches to content analysis on the internet. Whether in the form of memes, infographics or photographs, images stand to have a greater impact on users’ affective states, transferring information while also producing potentially visceral responses during otherwise innocuous everyday browsing. By developing a tool that speaks directly to the significance of affectively charged images, we propose a conception of both affects and users that is the opposite of the kind of sentiment analysis and extractivism that dominates corporate social media platforms. Specifically, the tool makes it possible for users to understand affect as layered, ambiguous and changing, rather than dealing with systems that either pin down or provoke an affective response to produce personalized recommendations or shape responses and behaviours. This tool does not aim to capture or freeze affect, but rather help users record and understand complex affective responses that would not become otherwise conscious. Combined with Image Flow, this tool enables new critical, exploratory and participatory methods to engage users in understanding their entanglements with information flows on social media.
