BACK

AwareSpace: Collaborative Exploration of Hiking Blogs with Surface Technologies


Mining useful information and making sense with blog data costs time and is usually difficult to organize findings. Text data like hiking blogs contain multiple facets and can be explored and understood from multiple directions. Existing natural language processing and visualization tools like topic modeling and word cloud provide ways to capture high-level semantics, but using these tools either limits directions to explore the information or requires the user to locate all useful information before being able to generate the visualization. When exploring the blog data, users need easy ways to annotate interesting information, summarize different aspects, and associate findings to generate knowledge structure. My project uses Surface technologies to support dynamic explorations of blog data: a tabletop computer and a vertical large display. The multi-touch tabletop provides a large interactive space to read, annotate and group blogs and the vertical display shows a connected word-cloud which reflects user’s interaction and draw hints on possible interesting knowledge. The surface connects the intuitive touch interaction with the machine learning technologies to help readers explore self-defined topics in the blog data. The system hides the complex machine learning algorithm from the user and provides an easy way to define, express and explore user-defined semantics by interaction and visualization.