Interactive data exploration using semantic windows

Abstract

We present a new interactive data exploration approach, called Semantic Windows (SW), in which users query for multidimensional "windows" of interest via standard DBMSstyle queries enhanced with exploration constructs. Users can specify SWs using (i) shape-based properties, e.g., "identify all 3-by-3 windows", as well as (ii) content-based properties, e.g., "identify all windows in which the average brightness of stars exceeds 0.8". This SW approach enables the interactive processing of a host of useful exploratory queries that are difficult to express and optimize using standard DBMS techniques.SW uses a sampling-guided, data-driven search strategy to explore the underlying data set and quickly identify windows of interest. To facilitate human-in-the-loop style interactive processing, SW is optimized to produce online results during query execution. To control the tension between online performance and query completion time, it uses a tunable, adaptive prefetching technique. To enable exploration of big data, the framework supports distributed computation.We describe the semantics and implementation of SW as a distributed layer on top of PostgreSQL. The experimental results with real astronomical and artificial data reveal that SW can offer online results quickly and continuously with little or no degradation in query completion times.

Overview of Data Exploration Techniques

“…Data prefetching has been studied within the context of data exploration for a number of query types such as multidimensional windows [36], data cubes [37,55,54] and spatial queries [63]. In this part of the tutorial we discuss these types of exploration queries and present alternative techniques for identifying promising data sets for pre-fetching, such as background execution of similar speculative queries [36,37] as well as indexing and searching past users' exploration trajectories [63].…”

“…Visualization tools for data exploration (e.g., [38,49,66]) are receiving growing interest while new exploration interfaces emerged (e.g., [18,32,45,57]) aiming to facilitate the user's interactions with the underlying database. In parallel, numerous novel optimizations have been proposed for offering interactive exploration times (e.g., [6,36,37]) while the database architecture has been re-examined to match the characteristics of the new exploration workloads (e.g., [8,27,28,39]). Together, these pieces of work contribute towards providing data exploration capabilities that enable users to extract knowledge out of data with ease and efficiently.…”

Efficient schemes for similarity-aware refinement of aggregation queries

“…These parameters are difficult to define especially for users without sufficient prior knowledge of the data. Another related work that supports query suggestion based on multiple aggregate constraints [14] has proposed a heuristic online search algorithm (HOSA) which aims to find a query that satisfies those aggregate constraints. The idea is to partition the search space into windows (using user specified parameter) then visit those windows in a decreasing order based on their utility (utility is how far a window's aggregate value is from the target one), and merge them as long as the utility increases.…”

Querying Geo-Textual Data

“…In contrast, Alexander et al [31] propose semantic window to study the region search problem for interactive data exploration of multidimensional data, in which a user explores a data space by posing a number of queries that find rectangular regions of interest. Region exploration.…”