A Practical Introduction to Geospatial Visual Analytics

Tutorial at GIScience 2010 in Zurich, Tuesday, the 14th of September, 2010

The tutorial introduces Visual Analytics – a new research discipline defined as the science of analytical reasoning facilitated by interactive visual interfaces. Visual analytics combines automated analysis techniques with interactive visualizations so that to extend the perceptual and cognitive abilities of humans and enable them to extract useful information and derive knowledge from large and complex data and to solve complex problems. Particularly, data and problems involving geospatial components and time are inherently complex and therefore call for visual analytics approaches.

We present a system of tools and methods for exploratory data analysis, including interactive maps, statistical graphics and multiple coordinated displays, data transformation methods for spatial and temporal data, and appropriate computational techniques such as cluster analysis and aggregation.

We focus on analyzing data about movement of discrete objects in geographical space. We consider three different examples of data:

movement of a single object during a long time period;
simultaneous movement of multiple objects, such as city traffic;
discrete (event-based) movement data, such as positions of Flickr photos.

The tutorial participants will be provided a complete version of Geospatial Visual Analytics Toolkit (a.k.a. CommonGIS) and several example real-world data sets. The toolkit may be used after the tutorial for research and educational purposes.

Tutorial organisers

Drs. Natalia Andrienko and Gennady Andrienko received Master degree in Computer Science from Kiev State University in 1985-1986 and Ph.D. equivalent in Computer Science from Moscow State University in 1992-1993. They undertook research on knowledge-based systems at the Institute for Mathematics of Moldavian Academy of Sciences (Kishinev, Moldova), then at the Institute for Mathematical Problems of Biology of Russian Academy of Sciences (Pushchino Research Center, Russia). Since 1997 Drs. Andrienko are research fellows at GMD, now Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS. They are authors of the monograph "Exploratory Analysis of Spatial and Temporal Data" (December 2005, Springer), 40+ peer-reviewed journal papers, 20+ book chapters, and 100+ papers in conference proceedings. They have been involved in numerous international research projects. Their research interests include geovisualization, information visualization with a focus on spatial and temporal data, visual analytics, interactive knowledge discovery and data mining, spatial decision support and optimization. Since 2007 Dr. Gennady Andrienko is chairing the ICA commission on GeoVisualization. He co-organized several scientific events on visual analytics, geovisualization, and visual data mining, and co-edited several special issues of journals and proceedings volumes.

Tutorial outline

Format of tutorial: 1/2 day

Introduction to visual analytics (30 minutes, lecture)
Introduction to visual analytics toolkit (60 minutes, installation, demo and hands-on), it's use for interactive geovisualization of different types of spatio-temporal data: events, time series, and trajectories of movement.
Analysis of movement (75 minutes, demo and hands-on)
- - pre-processing of movement data
- - detection of stops and other movement episodes
- - aggregation in space and time
- - progressive clustering of trajectories
- - visual summarisation of clusters of trajectories
Concluding remarks (15 minutes, overall discussion)
- - challenging {publicly available} data sets
- - open research problems, including computer science problems and privacy issues

Intended audience (introductory, advanced) and any background knowledge or skills required

The tutorial does not require special knowledge or skills.

Required materials (e.g. will you need internet access, lab access, specific software installed?)

The hands-on lessons require lab access (at least one standard PC for every two participants, it is possible to use individual MS Windows laptops of the participants) and installation of the visual analytics toolkit. The toolkit is implemented in Java, therefore an installation of a recent JRE is needed. A special web site with data, software and installation instructions will be prepared.

Internet access would be useful for accessing background maps from OpenStreetMap and Google. However, the tutorial can be done without internet connection as well.

Expected number of participants: 15-30