Making sense of wild data : using visualization to analyze in-the-wild video records

Abstract

In this paper we describe our use of information visualization to facilitate the analysis of in-the-wild video data. Video recording is often the method of choice when conducting in-the-wild studies. It results in highly rich and detailed data collections that can be revisited many times and analyzed from different perspectives. However, the qualitative analysis of video recordings collected in real-world settings is known as a tedious and time consuming activity, because the data can contain a large number of activity layers that have to be identified and manually extracted through video coding. We have utilized customized information visualizations to create visual representations of coded video recordings that consider particularly the temporal, social and spatial context of interactions. We describe how these visual abstractions from rich video data were valuable in various stages of our analysis process, including the cataloguing of video data, identifying research questions, in-depth analysis, and, finally, communicating our study results. We also point out various challenges that we identified in this process.