Sites Inria

Version française

ILDA Research team

Interacting with Large Data

Team presentation

In an increasing number of domains, computer users are faced with large datasets, that are often interlinked and organized according to elaborate structures thanks to new data models such as those that are arising with the development of, e.g., the Web of Data. Rather than seeing the inherent complexity of those data models as a hindrance, we aim at leveraging it to design new interactive systems that can better assist users in their data understanding and processing tasks.

These "Data-centric Interactive Systems" aim at providing users with the right information at the right time, presenting it in the most meaningful manner, and letting users efficiently manipulate, edit and share these data with others. This entails minimizing the effort required to retrieve and relate data from relevant sources; displaying data using presentation techniques that match the data's characteristics and the users' tasks; and providing users with means of interacting with the data that effectively support their train of thought.

Our emphasis is on users both as data consumers, using such systems to extract knowledge and insights from their data; and as data producers, using interactive systems to create, structure, transform and share data for consumption by others. Our long-term goal is to achieve what spreadsheets did for data in tabular form starting in the early 80's: to design a new generation of systems that have the potential to revolutionize how we interact with semantics-enriched, interlinked webs of heterogeneous data to help users transform raw data into actionable knowledge.

Research themes

  • Semantics-driven Data Manipulation. Our goal is to rethink how users manipulate large datasets, taking advantage of emerging technologies that give machine-processable semantics to, and enable the easy interlinking of, (semi-)structured data. Combined with novel input channels that feature a high level of expressive power, such as human gestures, we design interactive systems that enable more elaborate and efficient data manipulation.
  • Generalized Multi-scale Navigation. Beyond the manipulation of webs of data, an essential aspect of interacting with them is how to display them and let users efficiently navigate them, taking into account the fact that they are large and distributed over multiple endpoints. We want to generalize the concept of multi-scale navigation, applying it to interconnected, heterogeneous datasets.
  • Novel Forms of Input for Groups and Individuals. Analyzing and manipulating large datasets increasingly involves multiple users working together in a coordinated manner in multi-display environments (workstations, wall displays, handheld devices). Having a clear picture of what collaborators are doing is central in such situations. We aim at designing interactive systems that improve group awareness, by capturing distributed data-centric tasks and their semantics, studying them in particular situations, and designing interaction techniques that take advantage of novel input technologies such as tactile surfaces, 3D motion trackers and custom controllers built on-demand.

International and industrial relations

  • International Collaborations:
    • ALMA radio-observatory (Europe, North America, Japan)
    • INRIA Chile (Chile)
    • Cherenkov Telescope Array (Europe, Japan)
    • LSST (USA, Chile)
    • JAIST (Japan)
    • KISTI (South Korea)
    • Microsoft Research (USA)
    • Northwestern University (USA)
    • University of Konstanz (Germany)
    • University of Calgary (Canada)
  • National Collaborations:
    • IGN (ANR project MapMuxing)
    • BnF (Bibliothèque Nationale de France)
    • TKM
    • Telecom ParisTech

Keywords: Human-Computer Interaction; Interaction techniques for large datasets; Web of Data; Wall displays; Gesture-based interaction; Multi-user interaction.