Building data visualization and exploration support into an embedded database system: Big Data in the small

点击翻译按钮获取中文摘要

Relational database systems have been, and in many ways continue to be, the main driving engine behind thedata revolution. The hallmarks of these systems that enabled the revolution are relative efficiency, powerful,elegant, declarative query languages, and transactional consistency.Codd's Principle of Data Independence provided for a layer of abstraction that has allowed applications andusers to deal with the data at a logical level, without need to know how the database system deals with the dataat the physical level. Data independence allows for the database engine to ensure transactional consistency; thisremoves the need for application programs to do so. This bargain has made writing a wide range of large dataapplications significantly easier.But this bargain does come with costs too, of course. One is limited to what the "query-language" interfaces,such as SQL and Xquery, provide. While SQL has evolved into a quite competent platform for relational datamanagement and for complex analytic queries backing business applications, it is, with no small irony, poorlysuited for data exploration and visualization tasks. Meanwhile, exploration and visualization of large datacollections has become just as important a need today as has been traditional data management in the past.We plan to provide a scale-free, progressive, direct visualization of aggregate data for types of SQL queriestowards supporting data exploration.