Interactive analysis systems provide efficient and accessible means by which users of varying technical experience can comfortably manipulate and analyze data using interactive widgets. Widgets are elements of interaction within a user interface (e.g. scrollbar, button, etc). Interactions with these widgets produce database queries whose results determine the subsequent changes made to the current visualization made by the user. In this paper, we present a tool that extends IDEBench to ingest visualization interfaces and a dataset, and estimate the expected database load that would be generated by real users. Our tool analyzes the interactive capabilities of the visualization and creates the queries that support the various interactions. We began with a proof of concept implementation of every interaction widget, which led us to define three distinct sets of query templates that can support all interactions. We then show that these templates can be layered to imitate various interfaces and tailored to any dataset. Secondly, we simulate how users would interact with the proposed interface and report on the strain that such use would place on the database management system.

翻译：交互式分析系统为不同技术背景的用户提供了高效且易于访问的手段，使其能够通过交互式控件轻松操作和分析数据。控件是用户界面中的交互元素（例如滚动条、按钮等）。与这些控件的交互会产生数据库查询，其查询结果决定了用户对当前可视化所做的后续更改。在本文中，我们提出了一种扩展IDEBench的工具，该工具可接入可视化界面和数据集，并估算真实用户可能产生的预期数据库负载。我们的工具分析了可视化的交互能力，并生成了支持各种交互的查询。我们从每个交互控件的概念验证实现开始，进而定义了三种能够支持所有交互的不同查询模板集。然后，我们证明了这些模板可以分层组合以模拟不同的界面，并可针对任何数据集进行定制。其次，我们模拟了用户如何与所提出的界面进行交互，并报告了此类使用对数据库管理系统造成的压力。