Date: Tue, 05 Nov 1996 20:58:29 GMT Server: NCSA/1.5 Content-type: text/html Last-modified: Thu, 28 Sep 1995 17:30:31 GMT Content-length: 1931 NEXT! description

What is NEXT!?


The goal of the NEXT! project is to develop tools for exploratory analysis of large, non-traditional datasets. The name of the project is an acronym (Non-traditional data EXploration Toolkit) that is also a play on the fact that a major emphasis is on supporting *sequence* data exploration. We use the term *sequence* broadly, and include as examples trace data, time series data from a variety of domains, and satellite or sensor observations.

The NEXT! project has three main components. First, there is a visual data exploration tool called DEVise that supports a paradigm called *visual queries*. Each visual query is in essence an abstract description of how a collection of sequence data is to be mapped to the user's screen. The query serves as a template that can be applied to any data set with the appropriate schema, and the visual presentation can subsequently be explored in detail by standard operations such as scrolling, zooming and probing.

The second component of NEXT! is a system called SEQ that supports SQL-style queries over sequence data. SEQ is distinguished from standard SQL in that it provides easy (and efficient) ways to pose queries that use the sequentiality in the data set, for example, ``Find the N-day moving average of IBM''. SEQ can be used in conjunction with DeVise, with DeVise serving as a visual browser and SEQ as a data manipulation backend.

The third component of NEXT! is an on-going effort to investigate novel ways of analysing large data sets (often called *data mining*). In particular, we have developed a tool called BIRCH that can effectively cluster very large data sets.