Visual Digital Methods: Developing Approaches for the Analysis of Twitter Images
We’re excited that this panel from our ‘Picturing the Social’ research project was recently accepted for the 4th International Visual Methods conference, to be held in Brighton in September (16–18). Below are our panel abstract and the abstracts for the four proposed papers.
This panel engages with the conference’s ‘Exploring Digital Visual Methodologies’ strand. It reports findings from the ‘Picturing the Social: Transforming our Understanding of Images in Social Media and Big Data research’ project. This research is funded through an ESRC Transformative Research grant and has an ambitious focus, aiming to transform the social science research landscape by carving out a more central place for image research within the emerging fields of social media and Big Data research. The project is focused on better understanding the huge volumes of images that are now routinely shared on social media and what this means for society. This project involves an interdisciplinary team of seven researchers from four universities as well as industry with expertise in: Media and Communication Studies, Visual Culture, Software Studies and Sociology, Computer and Information Science. Through a series of smaller projects, informed by a facet methodology approach, it explores a number of issues, including how we might think about what a social media image is and how to study it.
The proposed panel addresses how we might account for the platform structures and affordances that underpin the circulation of Twitter images, the ways in which this platform allows users to share images and how this has changed over time, how we can study and understand the different kinds of images that are shared, and, finally how interdisciplinary work can open up new avenues for researching images on social media more widely.
The networked circulation of images on Twitter raises questions about how we might understand the nature and location of such images. It also raises questions about the agency involved in the generation and circulation of these images. Should the Twitter image and the online digital image in general be considered as something different from images distributed by earlier technologies because it is networked and constituted by code? How might approaches be developed to describe the visuality of the potential viewing of a Twitter image on multiple screens simultaneously? How might relationships between images on Twitter and their points of origination on and offline be considered? Moreover, given the networked relations of Twitter image circulation, the highly repetitive nature of some genres of Twitter image, and the algorithmic role of the platform itself, how can we think about authorship in terms of the Twitter image? The panel will explore these questions and suggest a range of strategies for developing approaches for the analysis of Twitter images.
Paper 1: Farida Vis (University of Sheffield) ‘A short history of image sharing on Twitter’
This paper explores image-sharing practices on Twitter during the platform’s 8-year history. Whilst it is possible to share images directly on the platform now, this has not always been the case. Image sharing was not possible when the platform was first launched in 2006. Strong interest from users, allowing them to add images to their tweets, gave rise to the development of a number of popular third-party applications and services, such as the now defunct Twitpic (launched in 2008). Over time Twitter has made image sharing both more possible and more restricted in different ways. Currently it shows images in users’ feeds only if these images have been uploaded directly to the platform. Images shared through other platforms and apps, such as Instagram, are shown as a link. This paper explores the implications of these different image sharing strategies from both a platform and user perspective, for example by looking at the ways in which images from Instagram have been treated over time, highlighting imaginative solutions users have developed to circumvent Twitter’s restrictions. In order to develop new digital visual methods for the study of social media images in this case for Twitter, we must also develop critical understandings of the dynamic intersections between platform, user and content.
Paper 2: Mike Thelwall (University of Wolverhampton) ‘Chatting through Pictures? A Classification of Images Tweeted in one week in the UK and USA’
Twitter is used by a substantial minority of the populations of many countries to share short messages, sometimes including images. Nevertheless, despite some research into specific images, such as selfies, and a few news stories about specific tweeted photographs, little known about the types of images that are typically shared. In response, this article reports a content analysis of random samples of 400 images tweeted from the UK and USA during a week at the end of 2014. Although most images were photographs, a substantial minority were hybrid or layered forms: phone screenshots; collages; captioned pictures; and pictures of text messages. About half were primarily of one or more people, including 10% that were selfies, but a wide variety of other things were also pictured. Some of the images were for advertising or to share a joke but in most cases the purpose of the tweet seemed to be to share the minutia of daily lives, performing the function of chat or gossip, sometimes in innovative ways.
Paper 3: Simon Faulkner (Manchester School of Art) ‘The uniqueness of Twitter images: Genre, location, and agency’
This paper seeks to reflect upon the uniqueness of Twitter images in relation to examples selected from a corpus of Twitter images circulated after the death of Margaret Thatcher in 2013. The paper will do this by considering three areas. First, it will consider different genre-types circulated on Twitter in comparison to those generated and distributed through earlier visual technologies. Second the paper will consider the question of the location of Twitter images in terms of their simultaneous existence on multiple screens and their repeated recirculation through acts of retweeting. Does the dynamic nature of Twitter images render them incomparable in terms of location to images disseminated by other pictorial technologies? Third, the paper will consider the role of agency within these processes. Given that Twitter images can be subject to multiple acts of spectatorship and retweeting within a network that has its own kind of agency, can we think about the agency in ways that are at all comparable to older image forms?
Paper 4: Francesco D’Orazio (Pulsar) ‘The pains, pleasures and myths of researching images on social media: exploring digital visual methodologies’
The semiotic complexity of images makes them trickier to study than words and without contextual qualitative understanding their analysis can prove misleading. The classic triad that has been the foundation of the visual culture approach (representation, meaning, culture) cannot be easily supported within a quantitative data-mining framework. How do we summon context, subjectivity and identify cultural codes through algorithmic approaches to mining images? Querying and sampling a representative dataset is also more difficult when dealing with images, as is storing the data, visualizing and manipulating it. Whilst text-mining tools address some of these issues above have been developed over the past 15 years, visual data mining is still where text mining was at the end of the 90s. More than that, when analyzing user generated textual content, concerns around ethics should be even higher when analyzing images, which are far more revealing than text. Setting the scene for the opportunities and the challenges around visual social media research, this paper looks at the new technologies and methods available to mine images today, the new cross-disciplinary approaches needed to make sense of images, the hybrid methods combining quantitative and qualitative analysis at scale and the key research use cases that visual social media research is enabling today and will be enabling over the coming years.
‘Picturing the Social: Transforming our Understanding of Images in Social Media and Big Data research’ is an 18-month research project that started in September 2014 and is based at the University of Sheffield in the United Kingdom. It is funded through an ESRC’s Transformative Research grant and is focused on transforming the social science research landscape by carving out a more central place for image research within the emerging fields of social media and Big Data research. The project aims to better understand the huge volumes of images that are now routinely shared on social media and what this means for society. This project involves an interdisciplinary team of seven researchers from four universities as well as industry with expertise in: Media and Communication Studies (Farida Vis and Anne Burns, University of Sheffield), Visual Culture (Simon Faulkner and Jim Aulich, Manchester School of Art), Software Studies and Sociology (Olga Goriunova, Warwick University), Computer and Information Science (Francesco D’Orazio, Pulsar and Mike Thelwall, University of Wolverhampton). The project is part of the Visual Social Media Lab.