Abstract:

The paper discusses "Reelchart", a visual interface for the exploration and analysis of collections of YouTube videos. The tool is meant to support "engaged research-led teaching" in the context of digital methods, defined as the entanglement of research, teaching and public engagement activities when doing academic work with and about the digital. Building on the notion of "programmed methods", where digital research methods and tools are often combined, we developed a tool and associated analytical technique for exploring video footage collected from Youtube. We envision the context of use of this tool in settings such as data sprints or teaching activities focused on studying social issues with the web in humanities, social sciences and design curricula.

After the visual turn in social media research, with the advent of video-only platforms such as TikTok and the adaptation of older platforms to increasingly promote videos over still images, video footage is now arguably the most available and shared type of content on the web. Although various tools and approaches for the analysis of collections of digital images have recently increased, there is still a lack of tools and analytical techniques catered explicitly to video footage. Due to their multimodal nature (including sound, music, still images, moving images, and text), videos present various layers of complexity that make their analysis particularly difficult. As of now, tools designed for video analysis require coding skills and focus on aesthetic characteristics such as colour, not allowing for a close reading of the video's content. To overcome these limitations, we developed "Reelchart", a visual exploration tool for the analysis of collections of YouTube videos. The interface offers a media visualisation approach for video footage. The tool is designed to blur the lines between quantitative and qualitative, allowing the analysis of video footage and its metadata (precisely words mentioned in the voiceover), keeping together video fragments and their context. Visualising videos through their very frames avoids data abstraction through graphical elements and helps keep the data points tethered to their original narrative context. Reelchart enables the analysis of two collections of Youtube videos and then uses the audio track to explore when and how various words are mentioned in the selected collections. Applying a media visualisation approach (as actual video frames are displayed, and the link to the actual videos is retained) allows for the contextualised reading of the data point. This aspect is particularly useful for videos, where semantics are conveyed through the interplay of audio, images and text, and thus the abstraction of any of those dimensions would hinder the reading of the data. The inclusion of segments following and preceding the selected one, and the ability to view the video with a single click, returns the selected word to the original narrative, suggesting a syntagmatic reading of the material. On the other hand, the display of videos as parallel sequences of segments encourages a paradigmatic reading of the analysed material, highlighting patterns emerging from repeated parts of the discourse. Blending quantitative and qualitative analyses, Reelchart fills a gap in the arsenal of tools employed in data sprints and teaching activities related to design, social sciences and the humanities, offering a no-code solution for the analysis of video material.

Keywords:

Digital methods, video, engaged research-led teaching, design, social media.