|Title / Titel||Visual Analytics of spatio-temporal gaze Point Patterns in Eye movements (PopEye)|
|Abstract (PDF, 14 KB)|
|Summary / Zusammenfassung||Anything that happens on Earth happens some place, at some time. Geographic phenomena exist within a dynamic multi-dimensional space-time continuum, at various levels of detail. Dynamic geographic phenomena can be generally conceptualized and represented as spatio-temporal patterns (e.g., trajectories of people or animals, flows of chemicals, or movements of eyes over maps), space-time processes (climate change, urban growth, human spatial cognition) or spatio-temporal events (e.g., earthquakes, winter Olympics, or human eyes fixating on a perceptually salient object in a scene). While current Geographic Information Systems (GIS) are excellent at representing, analyzing, and depicting spatially referenced data (and static snapshots of processes and events), they are still limited in handling temporally referenced spatial datasets such as dynamic patterns, processes, and events (Peuquet, 2002). Recent temporal Geographic Information Science (short: GIScience) research has focused on adding time to spatial databases (Hornsby and Egenhofer, 2000; Worboys, 2005). However, the development of spatio-temporal analytical tools and depiction methods (e.g., exploratory time-series analysis within a geographic framework) has lagged behind. In order to make GIS more effective for research in domains that focus on the explanation and prediction of dynamic patterns it is crucial to develop new analysis methods that can truly integrate the spatial and the temporal components of pattern analysis. This holds particularly true for the analysis of moving point entities in a dynamic spatial context.
Research objective: We propose to develop visual analytics methods and data exploration tools for the effective depiction and analysis of time-referenced spatial data sets at high resolution. We conceptualize time-dependent spatially referenced data as entities (e.g., humans, animals, money transactions, viruses, eye gaze locations, etc.) represented as points moving (changing x/y location) across a dynamic space context through time (t). The resulting visual analytics methods will be generic enough to be useful for all types of moving entity types, including various geographically relevant application domains (e.g., human trajectories and lifelines tracked with location based services, habitat analyses with tagged animals in wildlife biology, eye movements analyses in geographic visualization experiments, etc.). As a generic data proxy for moving points we utilize self-collected gaze point samples from eye movement experiments on spatial displays.
Research questions leading this effort can be seen along two tracks:
• How can we efficiently discover, extract knowledge and communicate spatio-temporal information buried in large collections of time-referenced spatial data at high temporal resolution (e.g., eye gaze data), within a dynamic spatial context? Specifically how can we detect, analyze and depict the order of (eye movement) events?
• How can we efficiently mine time dependent spatial data patterns and effectively communicate results to humans with limited spatio-temporal processing capabilities? Specifically, how can we summarize large amounts of (gaze) point movement data? How can we discover recurring patterns or surprising patterns?
This proposed research is based upon previous research aimed at developing novel visual analytics methods for exploring moving point objects at the individual level and for groups. Drawing upon and extending the previously developed techniques we propose to transfer knowledge and deepen the integration of spatial and temporal exploratory data analysis methods currently being developed in complementary research fields such as geographic information visualization and spatialization, Temporal GIS, as well as spatial and temporal data mining.
This 3-year research program is organized into two interleaved tracks: one line of research will emphasize on the analytics part (spatializations, temporalization, transformations, etc.), while the second will focus on exploratory visualization, interaction and evaluation. Key outputs are expected along three cyclic work phases for each track, including 1) interdisciplinary theoretical contributions, 2) technological components, and 3) empirical evaluation results. First, knowledge integration across complementary research fields (as mentioned above) is carried out at the theoretical level (spatialization methods of time-referenced data and temporalization approaches of spatially referenced data sets), 2) proof-of-concepts are implemented, and 3) proof-of-concepts (methods and tools) are evaluated with users (empirical evaluation of usefulness and usability).
|Keywords / Suchbegriffe||Temporal GISystems, space-time integration, spatio-temporal pattern analysis, moving point entities, eye movement tracking, gaze point patterns|
|Project leadership and contacts /
Projektleitung und Kontakte
|Funding source(s) /
|SNF (Personen- und Projektförderung)
|In collaboration with /
In Zusammenarbeit mit
|Duration of Project / Projektdauer||Mar 2007 to Sep 2009|