Where: CSSSI 24-Hour Space
When: Friday, October 18th; 11.30a for the food, 12p for the talk
Emanuel Zgraggen, PhD Student at the Brown University Computer Science Department, will present his work on interactive data visualization. In particular, Emanuel will discuss PanoramicData, his adaptation of the Brown University WorkTop project that focuses on information interfaces, multimedia, and text-driven scholarship.
PanoramicData
Data is omnipresent in today’s world. Analyzing and understanding data plays an important role in a variety of different domains. Imagine an investment banker who tries to decide which stock to buy next and who backs up her decision based on historical stock-prices, a web developer looking at server log files to figure out which parts she should be optimizing, or a sports fan who tries to explain to his friends, using game statistics, that their favorite team plays better if player X is not on the field. Typically, users currently rely on domain-specific tools or websites (e.g., E*TRADE, Google Analytics, NBA Stats website) to accomplish such tasks. For simplicity, such domain-specific applications provide customized tools that make it easy to perform common queries but do not allow for general data probing. Tools like Excel or SQL provide for more general analysis but require significant programming knowledge.
We introduce PanoramicData, an interactive system for exploring, visualizing, and analyzing datasets. It attempts to provide the full power of SQL in an easily approachable user experience. PanoramicData allows familiar rich data visualizations, such as charts and table views, to be linked together into data flow networks. The rich focus+context these data panoramas provide facilitates non-linear workflow in which each component can be changed individually (focus) while viewing its propagated effects on connected parts (context). The resulting picture can be thought of as data narrative — an interactive panorama depicting the customized analysis of a dataset.
WorkTop
WorkTop was designed to enhance the ability of students (and scholars at all levels) to more efficiently perform fundamental scholarly tasks while at the same time providing new opportunities for collaborative learning.
Driven by the success of Integrated Development Environments (IDEs) for supporting recurring and complex code development tasks, we have been researching a similar platform for small workgroup, document, and multimedia-oriented workflows.
Research-driven tasks, whether performed by students or seasoned scholars, involve frequent transitions across a broad spectrum of document-centric activities, ranging from searching and viewing to writing and publishing. Despite technological advances, unnecessarily inefficient and complicated transitions plague humanities workflows such as converting between document formats, using small windows to view large document collections, and coercing free-form spatial mental models and visual data into rigid, generally lexical structures.
WorkTop addresses these and similar issues by reducing the technological barriers to capturing, displaying, and linking heterogeneous documents or document fragments within a unified workspace. Keywords can be easily applied to entire documents or specific regions. Fine-grained, typed, and bidirectional hyperlinks can be created between images, videos, notes, PDFs, and webpages. Rich text annotations can be added as easily as placing a sticky note.
Additionally, WorkTop provides a variety of structured workflow assistants, such as templates for publishing collected materials as stylized multimedia web pages, search operations which can be persisted as smart folders, and snapshots that capture the workspace state so that tasks can be seamlessly resumed at a later time.
Finally, WorkTop supports small-team collaboration by capturing all work in a replicated, shared database. We are also in the process of extending WorkTop to support multitouch and stylus input so that it can serve better as a live presentation tool on electronic whiteboards and tablets.