UCLA

The proliferation of sensors in consumer mobile phones has made them a powerful data collection platform. We have become more interested in collecting data to document and learn about our own lives, our relationships with the community, and the environment we live in. However, current research is mostly focused on analysis of quantitative data, and a lot of important information that can be extracted from qualitative data such as images and text annotations is being ignored. The focus of this dissertation was to research and build techniques based on narratology to help users analyze and make sense of the qualitative data they have collected and to communicate the information and knowledge they extract from the data though narratives. More specifically, our main contribution was in the creation and evaluation of two techniques that were fundamental to narration tools: clustering and conflict discovery. We used these techniques to build the Narreme toolset that enables user-assisted creation of compelling narratives with participatory sensing data. The toolset consists of three tools: (a) Narreme clustering tool which helps a user organize her thoughts and identify supporting qualitative data consisting of images and text for her narratives, (b) Narreme conflict discovery tool helps the user better understand where conflict could be introduced in her narratives, and, (c) Narreme video editing tool which is used to create a video slideshow narrative and adjust the order based on a narrative ten- sion curve. We evaluated the effectiveness of our techniques and toolset through user studies.