We propose a new visualization interface for navigating the ACM digital library. It supports effective literature exploration with a set of web-based functions including search, detail summary, conference summary, author summary, and citeology. These functions are designed and integrated with enhanced perceptual understanding and human-machine interaction, where colors, diagrams, layouts and other informative visualization factors are utilized to analyze the collective metadata from the ACM digital library.
We use a large-scale data set of the titles, authors, categories and abstracts based on the ACM digital library. A phrase extraction algorithm is designed to retrieve meaningful phrases from the data set. All the web based functions mentioned above uses this algorithm. These phrases, instead of single keywords, can represent the publications with improved semantics, which enhance the visualization output and user experience. We do this by sequentially scanning the paper’s abstract and using pre-defined dictionaries, punctuations to pull out groups of meaningful phrases and throwing out junk words such as verbs, possessive pronouns and other pre-defined stop words that would prevent us from returning meaningful data. The visual interface provides an advanced platform for researchers in their literature study. It can be further extended to the exploration of other libraries and databases.