Navigation auf uzh.ch

Suche

Department of Informatics Interactive Visual Data Analysis Group

Interactive Visual Search and Exploration for Digital Libraries

A digital library (DL) is a database of digital objects (documents, items) that can include text, still images, audio, video, digital documents, or other digital media formats accessible through the internet. Digital library systems (DLS) enable users to gain access to digital objects (documents, items) through interactive search and filtering capability. Beyond that, DLS may also support exploration and data analysis capabilities to enable users gaining insights even while browsing for interesting digital objects. This is where methods and techniques developed in IVDA can help most.

Our research focuses on changing paradigms on how digital libraries and digital library systems will support future users searching for digital documents, items, and objects of any kind. Examples include various types of textual documents, research data, multimedia data, shopping items, and more. The vision is to enhance today's digital library services with methods and techniques developed in the realm of IVDA. Thus, main contributions will be to transfer technology into a data-intensive environment that, however, lacks sophisticated computer science and data science capability. Concrete research directions include:

  • Enhanced interactive and visual search and filtering techniques
  • Data exploration and knowledge generation through DLs
  • Content-based access to complex digital objects
  • Search for similarity measures that would be most useful
  • Human-centered AI in digital libraries
  • Human-model teaming of digitalization processes
  • Designing novel interactive visual digital library systems

[Paper] VisInfo: A Digital Library System for Time Series Documents

personalizedMusicClassifier
"Visual Catalog" for the exploration of time series patterns. An overview of 125,000 daily temperature curves. Each of the 80 cells represents about 1,500 daily curve patterns of the data set (see small blue numbers)

VisInfo is a web-based digital library system (DLS) with the goal to provide visual access to time series research data. Based on an exploratory search (ES) concept, VisInfo at first provides a content-based overview visualization of large amounts of time series research data. Further, the system enables the user to define visual queries by example or by sketch. Finally, VisInfo presents visual-interactive capability for the exploration of search results. The development process of VisInfo was based on the user-centered design principle. Experts from computer science, a scientific digital library, usability engineering, and scientists from the earth, and environmental sciences were involved in an interdisciplinary approach. We report on comprehensive user studies in the requirement analysis phase based on paper prototyping, user interviews, screen casts, and user questionnaires. Heuristic evaluations and two usability testing rounds were applied during the system implementation and the deployment phase and certify measurable improvements for our DLS. Based on the lessons learned in VisInfo, we suggest a generalized project workflow that may be applied in related, prospective approaches.

visinfo2
List-based interface showing the result of a time series search (retrieved time series patterns are highlighted red). Search result visualization including metadata facets (left), list-based visualization of the data content (center), and respective metadata (right). The result set corresponds to a query that was formulated by a user (query by example or by sketch).

Materials and Bibliographic Data

[Project] Tag-Xplore: Interactive Exploration of Annotation Practices in Digital Editions

With Tag-Xplore, we design and develop an interactive data exploration system that enables curators of digital editions to understand commonalities and differences in the use of metadata tags in multiple digital editions, thus enhancing standardization processes.

Materials and Bibliographic Data

  • External Project Website: DSI Tag-Xplore
  • Prototype Demo (accessible from UZH network): Demo 

[Project] ORD-Xplore: Bridging Open Research Data Collections

ORD-Xplore is a project that aims to support the Open Research Data (ORD) standards and FAIR principles for repositories of digital editions to better support accessibility and usability.

Materials and Further Links:

  • External Project Website: DSI ORD-Xplore
  • Prototype Demo (accessible from UZH network): Demo