Visualizing Email Content: Portraying Relationships From Conversational Histories Fernanda B. Viegas Scott Golder Judith Donath Speaker: Yi-Ching Huang 2007/09/04
About this paper • Fernanda B. Viégas , Scott Golder , Judith Donath, Visualizing email content: portraying relationships from conversational histories, Proceedings of the SIGCHI conference on Human Factors in computing systems, April 22-27, 2006, Montréal, Québec, Canada
Email Visualization • Thread-based visualizations • Social-network visualizations • Temporal visualizations • Contact-based visualizations
• They are useful for showing the networks of acquaintanceship and the temporal rhythms of interactions, but...
• They do not provide any clues about the topics people discuss or the type of language they use with different members of their social circles.
Themail
Themail • Themail visualize an individual’s email content over time • show keywords along a timeline • keywords have different colors and sizes
Two main questions • What sorts of things do I talk about with each of my email contacts? • How do my email conversations with one person differ from those with other people?
• Multiple layers of information • Yearly words : reveal the most used terms over an entire year of email exchange • Monthly words : are the most distinctive and frequency used words in email conversation over a month
How does it work?
Themail • Themail processor • merge multiple account • calculate topic words • TFIDF algorithm • Themail visualization • interacting with Themail • adjusting the time scale
Our Idea: Bookmark Rhythms
Main Idea • Show an individual’s interest change over time • Visualize the rhythm of interest change
Our Approach • Data source from • social bookmarking website: del.icio.us • Extract multiple layer information • overall topic • topic exchange over time
Information Processing • Clustering • Classification • Semantic v.s. Color analysis
Visualization • overall visualization • tag-cloud (color & size) • temporal visualization • rhythm (timeline) • interact with data
Milestone • 9/4 ~ 9/11 • visualize the raw data • implement ConceptNet
Recommend
More recommend
