作者: Nikos Bikakis , John Liagouris , Maria Krommyda , George Papastefanatos , Timos Sellis
DOI: 10.1109/ICDE.2016.7498340
关键词:
摘要: We present a novel platform for the interactive visualization of very large graphs. The enables user to interact with visualized graph in way that is similar exploration maps at multiple levels. Our approach involves an offline preprocessing phase builds layout by assigning coordinates its nodes respect Euclidean plane. respective points are indexed spatial data structure, i.e., R-tree, and stored database. Multiple abstraction layers based on various criteria also created offline, they similarly so can explore dataset different levels granularity, depending her particular needs. Then, our system translates operations into simple efficient (i.e., window queries) backend. This technique allows fine-grained access graphs extremely low latency memory requirements without compromising functionality tool. web-based prototype supports three main operations: (1) navigation, (2) multi-level exploration, (3) keyword search metadata.