Christophe Hurter is with DSNA/DTI R&D, ENAC and IRIT/IHCS. christophe.hurter@aviation-civile.gouv.fr
Benjamin Tissoires is with DSNA/DTI R&D, ENAC and IRIT/IHCS. tissoire@cena.fr
Stéphane Conversy is with ENAC and IRIT/IHCS. stephane.conversy@enac.fr.

In this paper, we describe a set of visualization methods with an accumulation tool to perform interactive data exploration. Accumulation maps or Kernel Density Estimation (KDE) maps, count the amount of data accumulated at a certain location. The accumulation tool addresses the cluttering issues when displaying large amounts of data. But the accumulation tool can also be used to unveil patterns, to detect outliers and flaws in datasets. Through these applied examples, we show how the accumulation tool and its real time applications take advantage of human vision and are therefore assets for data exploration and validation. As the accumulation tool uses GPU techniques, it can be used in real-time with large datasets.

Christophe Lienert, Institute of Cartography of ETH Zurich, and the Oeschger Centre for Climate Change Research of the University of Berne, E-Mail: lienert@karto.baug.ethz.ch (main and corresponding author)
Bernhard Jenny, Institute of Cartography of ETH Zurich , E-Mail: jenny@karto.baug.ethz.ch
Rolf Weingartner, Geographical Institute, and the Oeschger Centre for Climate Research of the University of Berne, E-Mail: wein@giub.unibe.ch
Lorenz Hurni, Institute of Cartography of ETH Zurich , E-Mail: hurni@karto.baug.ethz.ch

Hydrological decision support systems are essential for early warning and real time management of flood events. Hydrologists require an integrated system that provides all temporal information and tools relevant for decision making. Current hydrological decision support systems predominantly visualize data by charts and tables, and do not take advantage of digital cartography. Furthermore, information is often dispersed on various analogous and digital media, and their non-automated compilation is time-consuming. This paper presents a real time decision support system for flood monitoring and water resources management that (1) automates the collection and archiving of a multitude of real time data; (2) uses an interactive and animated map as a central node for data access and visualization; (3) visualizes and filters statistical surfaces in 3D for the quick localization extreme values; and (4) provides temporal navigation for comparing the momentary hydrological situation with archived events, in order to estimate the further development. The described system for hydrological decision support has been developed in close collaboration with hydrologists working in operational flood management, and has received encouraging feedback.

Werner Scheltjens
Navigocorpus post-doctoral fellow Ecole Normale Supérieure de Lyon 15, Parvis Descartes 69007 Lyon France

In this paper, an attempt will be made to establish a connection between exploratory and analytical approaches to the visualization of movement data and the general analytical framework of economic evolution with the final aim of advancing historical analysis. It is expected that the proposed connection will be beneficial for both method, theory and economic history. It will be argued that the integration of methods for the study of movement data in the general analytical framework of economic evolution provides useful directions for the visual analysis of movement data, enhances the applicability of the general analytical framework to historical cases and offers the economic historian a comprehensive framework that includes both theory and method. The scientific argument of this paper will be substantiated by the analysis of a historical case on Scandinavian shipping in 1784-1795.

Petr Glos, Institute of Computer Science, Masaryk University, Botanicka 68a, 602 00 Brno, Czech Republic, E-Mail: glos@ics.muni.cz
Luboš Popelínský, KD Lab, Faculty of Informatics, Masaryk University, Botanicka 68a, 602 00 Brno, Czech Republic,E-Mail:popel@fi.muni.cz

Masaryk University maintains a digital version of building passport that currently consists of approximately 200 buildings and 17,000 rooms. For data and process visualization ESRI ArcGIS has been successfully used. However, for deep understanding of running processes in a room, namely for anomaly detection, we would need an analytic tool that can predict (or at least detect) such rare events. A rare event is a pattern that is rare but important for precaution. Such event is e.g. fire (or an increase of a temperature), fast repeated switching on/off of a device, or water pipe disruption. In this paper we focus on anomaly detection in temperature trends. We first describe shortly the building management system and then we overview visualization tools that are currently used. In the second part of this text we introduce an analytic method that serve for detection of anomalies – unexpected behaviour in temperature – in a room. As the number of temperature measurement varies (from less than 90 to more than 80000) depending on a way of measurement and temperature changes in a place, various aggregated values are first computed. This method combines visualization with automatic classification and outlier detection.

C. Pöelitz and G. & N. Andrienko
University of Bonn & Fraunhofer IAIS, Germany

The paper deals with density-based clustering of events, i.e. objects positioned in space and time, such as occurrences of earthquakes, forest fires, mobile phone calls, or photos taken by Flickr users. Finding concentrations of events in space and time can help to discover interesting places and time periods. The spatial and temporal properties of event clusters, in particular, their spatial and temporal extents and densities, can be related to each other. According to Tobler’s Law, the relationship can be described as follows. Events in a small area can be sparse in time and still connected. On the other hand, events in large areas are likely to be connected if they are close in time. Hence, the temporal distance threshold for density-based clustering should vary depending on the spatial extent of the area in which events happen. Therefore, we suggest a two-step clustering method. In the first step, spatial clusters of events are detected. In the second step, density-based clustering is applied to the temporal positions of spatially clustered events. The temporal distance threshold is chosen individually for each spatial cluster depending on its spatial extent. We demonstrate the work of the method on several examples of real data.

John Doyle Strategic Research cluster in Advanced Geocomputation at the National University of Ireland Maynooth, Maynooth, Co.Kildare, Ireland. E-mail:jcdoyle@eeng.nuim.ie
Seán McLoone Strategic Research cluster in Advanced Geocomputation at the National University of Ireland Maynooth, Maynooth, Co.Kildare, Ireland.
Tim McCarthy Strategic Research cluster in Advanced Geocomputation at the National University of Ireland Maynooth, Maynooth, Co.Kildare, Ireland.
Ronan Farrell Strategic Research cluster in Advanced Geocomputation at the National University of Ireland Maynooth, Maynooth, Co.Kildare, Ireland.

Methods for observing the dynamics between people and the environments they pass through are becoming ever more increasingly easier to observe. Recent research has indicated that these dynamics can be observed and interpreted on urban scales, through cellular mobile device activity patterns. This paper presents a national scale visualisation of cellular mobile device activity. Results allude to the rural-urban daily migration patterns evident in the activity metrics present in the commuter belt region of Dublin.

Arzu Coltekin*, Sara I. Fabrikant* and Martin Lacayo**
* Geographic Information Visualization and Analysis Group (GIVA), Department of Geography, University of Zurich, Switzerland, E-mail: name.lastname@geo.uzh.ch
** Department of Geography, San Diego State University, USA, E-Mail: lastname@rohan.sdsu.edu

In the background of this study lies an experiment primarily concerned with user experience of map designs, in which 30 participants solved three typical map use tasks over two interfaces that utilized identical information. During this experiment, along with a traditional usability setup, viewers' eyes were tracked. Obtained gaze data allows immediate qualitative insight into the visual search processes, as well as some quantitative (statistical) inferences that are otherwise not possible. In this study, as a follow up to map interface comparison experiment, we analyze a selected set of eye movement recordings exploring user strategies for efficiency. Do faster (more efficient) participants in fact display a pattern different than others, and if so, can a cartographer (designer of a map) learn from this? In trying to answer this question, we explore the gaze data with a temporal focus, that is, we analyze fixation sequences using our software Point Pattern Analyst in combination with several other software packages for different stages of the process.

Jacqueline Young, Niklaus Ashton, Claus Rinner
Department of Geography, Ryerson University

Population health is influenced by many socioeconomic and demographic factors that may include levels of employment, income, education, ethnicity and age. For health planning and service delivery, it is important to take into account demographic trends over time. This temporal component is usually incorporated into analyses by comparing multiple maps of variables at different points in time. In this study demographic variables with spatial and temporal components are used in a multi-criteria analysis within an interactive spatial decision support tool. We illustrate how the exploration of an area-based composite index over time can help analysts with identifying trends of increasing social deprivation and health-care needs. The paper focuses on the conceptual challenges of spatio-temporal multi-criteria analysis due to changing geographic boundaries, the standardization of variables across time, comparability of variables, and comparability of index scores.

Xia Li, Menno-Jan Kraak
Xia Li is with ITC, University of Twente , E-Mail: Xia@itc.nl
Menno-Jan Kraak is with ITC, University of Twente , E-Mail: Kraak@itc.nl.

abstract?

Jim Thomas
AAAS, PNNL Fellow; Senior Science Advisor National Visualization and Analytics Center, Pacific Northwest National Laboratory; President and CEO DiscoverVisualAnalytics LLC

Urška Demšar (1) and Kirsi Virrantaus (2)
(1) National Centre for Geocomputation, National University of Ireland Maynooth, Ireland, urska.demsar@nuim.ie
(2) Geoinformatics and Cartography, Department of Surveying, School of Science and Technology, Aalto University, Espoo, Finland, kirsi.virrantaus@tkk.fi

Data about movements of objects are often collected as trajectories in space and time. In cartography, a typical way to visualise and explore such data is to use a space-time cube, where trajectories are shown as 3D polylines through space and time. With larger and larger movement datasets, this display quickly becomes overcluttered and unclear. In this paper we introduce the concept of a 3D space-time kernel density of trajectories to approach the problem of overcluttering the space-time cube. The space-time density is a generalisation of standard kernel density around 2D point data into 3D density around 3D polyline data (i.e. trajectories). We present the algorithm for space-time density, test it on simulated data and show some basic visualisations of the resulting density volume. We also present an application on real-time movement data (vessel movement trajectories in Helsinki harbour) and discuss the potential of the method for improvement of the current maritime surveillance system in Finland.

Niels Willems, Visualization group at Eindhoven University of Technology
Willem Robert van Hage, Web & Media group at VU University Amsterdam
Gerben de Vries, Theoretical Computer Science group at University of Amsterdam
Jeroen Janssens, Tilburg Centre for Creative Computing at Tilburg University
Véronique Malaisé, Web & Media group at VU University Amsterdam

We present an integrated and multi-disciplinary approach for analyzing behavior of moving objects. The results are ongoing research of four different partners in the Dutch Poseidon project [http://www.esi.nl/poseidon] where we aim for new developments for Maritime Safety and Security (MSS) systems to monitor vessels. We focus on the following requirements for an MSS system: abstraction of large amounts of trajectory sensor data, fusion of multiple heterogeneous data sources, and visual analysis of the combined data sources. We start by extracting segments of consistent movement from trajectory data, which we store as instances of the Simple Event Model (SEM), an event ontology represented in the Resource Description Framework (RDF). Then we add data from the web about vessels to enrich the sensor data. This additional information is integrated with the representation of the vessels (actors) in SEM. The enriched trajectory data is stored in a knowledge base, which is queried by a visual analytics tool to search for spatio-temporal patterns. Although our approach is dedicated to MSS systems, we expect it to be useful in other domains.

Professor Mikael Jern, NCVA - National Center for Visual Analytics http://ncva.itn.liu.se ITN/VITA Linköping University, S-601 74 Norrköping, Sweden, E-Mail: mikael.jern@itn.liu.se

Official statistics such as demographics, environment, health, social-economy and education from national and sub-national sources are a rich and important source of information for many important aspects of life and should be considered to be more used and acknowledged in education. Educators and their students would be able to get informed and at the same time participate in increasing the knowledge on how life is lived and can be improved. Public statistics databases, e.g. EuroStat, OECD, Worldbank, WHO and numerous national statistics bureaus etc. can be reached on the Internet.
A better understanding of how educators and their students can elicit better user understanding and participation by exploiting dynamic web-enabled geovisual analytics and its associated science of perception in learning is the focus of this presentation in relation to the use of multidimensional spatio-temporal statistical data. Public available Web tools (Open eXplorer) are explained that help and engage educators to communicate progress initiatives measuring economic, social, educational, health and environmental developments to students and citizens. NCVA has since 2008 in close research collaboration with OECD developed and evaluated geovisual analytics tools for exploring and communicating statistical information. Storytelling and publishing statistics news in blogs or digital newspapers are examples of our latest research direction. Means are explained how the author (educator) 1) select spatio-temporal and multidimensional national or sub-national statistical data, 2) explore and discern trends and patterns, 3) then orchestrate and describe metadata, 4) collaborate with colleagues to confirm and 5) finally publish essential gained insight and knowledge embedded as dynamic visualization “Vislet” in blogs or web pages with associate metadata. The author can guide the reader in the directions of both context and discovery while at the same time follow the analyst’s way of logical reasoning. We are moving away from a clear distinction between authors and readers affecting the process through which knowledge is created and the traditional models which support editorial work. Value no longer relies solely on the content but also on the ability to access this information. Audiences are increasingly gathered around Web enabled technologies and this distribution channel is, more than ever, in control of the information value chain.

Yi Qiang, Katrin Asmussen, Matthias Delafontaine, Steven Logghe, Philippe De Maeyer, Nico Van de Weghe
1. Department of Geography, Ghent University Krijgslaan 281, Ghent 9000, Belgium {yi.qiang, katrin.asmussen, matthias.delafontaine, philippe.demaeyer, nico.vandeweghe}@ugent.be
2. Be-Mobile NV Technologiepark 12b 9052, Ghent- Belgium steven.logghe@be-mobile.be www.be-mobile.be

This paper introduces an innovative representation of time series data, namely, the Continuous Triangular Model (CTM). In CTM, time series data of different intervals are displayed by a continuous field. Every position in the field represents a specific time interval. From the bottom to the top, the length of intervals increases. The value at each position represents a certain kind of information during the interval. In this paper, we show that CTM provides a direct overview of time series data at multiple scales. On the other hand, sets of intervals that are decided by values in the third dimension can be represented by areas in the CTM field, forming a basis for answering queries with non-temporal constraints. Besides time series data, CTM can also be applied to other types of sequential data. Since CTM represents data in a two-dimensional space, a lot of techniques in geographical information science are available to manipulate and analyse CTM.

Keywords: Continuous Triangular Model, Triangular Model, Time Interval, Time Series Analysis, Geographical Information Science

Doris Dransch, Patrick Köthur, Sven Schulte, Volker Klemann, and Henryk Dobslaw
Helmholtz Center Potsdam, German Research Centre for Geosciences
University of Applied Science Anhalt

Frank Hardisty and Alexander Klippel
Department of Geography 302 Walker Building and John A. Dutton e-Education Institute 2217 Earth & Engineering Sciences Building, The Pennsylvania State University, University Park, PA, 16801, USA

Many interesting analysis problems (for example, disease surveillance) would become more tractable if their spatio-temporal structure was better understood. Specifically, it would be helpful to be able to identify autocorrelation in space and time simultaneously. Some of the most commonly used measures of spatial association are LISA statistics, such as the Local Moran’s I or the Getis-Ord Gi*, however these have not been applied to the spatio-temporal case (including many time steps) due to computational limitations. We have implemented a spatio-temporal version of the Local Moran’s I, and claim two advances. First, we exploit the fact that there are a limited number of topological relationships present in the data to make Monte Carlo simulation computationally tractable, and thereby bypass the “curse of dimensionality”. Second, we developed a tool (LISTA-Viz) for interacting with the spatio-temporal structure uncovered by the statistics.

Cyril Faucher*, Cyril Tissot**, Jean-Yves Lafaye*, and Frédéric Bertrand*
*: L3i University of La Rochelle, France
**: UMR 6554 CNRS, France
{cyril.faucher@univ-lr.fr, cyril.tissot@univ-brest.fr, jean-yves.lafaye@univ-lr.fr, frederic.bertrand@univ-lr.fr}

Within a decision process about space and time data, visualization may be seen as twofold. On the one hand, visualization reflects a synthetic point of view and expresses semantic aspects of the current results; on the other hand, it stands as a support for the analyst’s intuition and suggests novel hypothesis or further computations. In any case, visualization exploits data and knowledge that should be recorded somewhere. Often, large and possibly high dimensional sets of data are to be processed for analysis, reasoning and decision making. It is clear that time datasets made of calendar elements (concrete dates) hide a great deal of the semantics that can be extracted by ad hoc methods and expressed in natural or specific languages. In this paper, we provide a general UML object model for specifying temporal knowledge. We of course account for usual time stamps and durations referring to sets of calendar data, but also focus on specifying concise representations for many kinds of periodic events - somewhat in the same way the Fourier transform can compress signals - except that our approach keeps close to the natural language and to the domain model. However, one important point is that in contrast with natural languages, our specification language is unambiguous. This paper presents a contribution about a use case dedicated to providing time specification facilities for a Multiagent simulation System applied to the modelling of marine fishing activities. The final goal is to provide the user with a language that handles a large set of abstract temporal expressions, especially including periodic ones.

S. Hadlak, C. Tominski, H.-J. Schulz, and H. Schumann
University of Rostock A.-Einstein-Str. 21 D-18059 Rostock Germany

While there are several techniques addressing specific aspects of spatio-temporal visualization, approaches that cope with space, time, data, AND structure are rare. With this paper we take a step to fill this gap. By combining various well-established concepts we achieve a reasonably complete visualization of all of the aforementioned aspects, where our focus is on hierarchical structures. We embed hierarchies directly into regions of a map display using a novel variant of a point-based layout. Layering and animation are applied to visualize temporal aspects. Depending on analysis goals, users can switch between representations that emphasize data attributes or hierarchical structures. Interaction techniques support users in navigating the data and their visualization.

Itzhak Omer, Peter Bak and Tobias Schreck
Itzhak Omer is with Tel Aviv University, E-Mail: omery@tau.ac.il
Peter Bak is with University of Constance, E-Mail: bak@dbvis.inf.unikonstanz.de
Tobias Schreck is with University of Darmstadt, E-Mail: tobias.schreck@gris.informatik.tu-darmstadt.de

Our research investigates the dynamics of ethnic residential pattern in the city based on geo-referenced house-level socio-demographic and infrastructure data. We base our research on the spatial diffusion model and on visual analytic methods to enable identification, presentation and mapping local land-use spatial configurations that have differential effect on the dynamics of that pattern. The research was instantiated on the Arab community residential pattern in Jaffa - Israel. The data on residential distribution was collected for more than forty years at four different time moments during the population and housing censuses of the Central Bureau of Statistics and Ministry of Interior. For safeguarding privacy, we applied aggregation and transformation on the representation level, but conducted our analysis on the raw data. Through this investigation, we have identified and explained the effects of spatial land-use configurations on the ethnic composition change-rate of the residence in analytical and objective terms, and in the same time safeguarding sensitive house-related details.

Ms.C. (tech.) Salla Multimäki,
Geoinformatics and Cartography, Department of Surveying, School of Science and Technology, Aalto University, Espoo,Finland, e-mail: salla.multimaki@tkk.fi

This abstract presents ongoing research in the field of the visualization of spatio-temporal information. The case study is related to the maritime surveillance system with a real-time situation picture that is used by the Finnish authorities. The application presented here is a tool that combines real-time spatio-temporal information from several sources, presents them on a map and shares the information between different authorities. The aim of this study is to reveal the diverse requirements of various user types and thus show that one non-customized application cannot serve different users in the best possible way, even when the surveillance data are originally the same

Ulanbek D. Turdukulov, Connie A. Blok, Barend Köbben and Javier Morales
Department of Geo-Information processing, Faculty of Geo-Information Science and Earth Observation (ITC), University of Twente. E-mail: {turdukulov, blok, kobben, morales}@itc.nl

Geographic information technologies are evolving from stand-alone systems to a distributed model of independent web services. In parallel, voluminous geographic data are being collected with modern data acquisition techniques such as remote sensing and personal navigation devices. There is an urgent need for effective and efficient methods to integrate and explore relationships between remote sensing and trajectory datasets. Examples where integration is needed are trajectories of hurricane and iceberg movements that need to be coupled with temperature images to study climate change; wild animal trajectories need to be integrated with vegetation conditions to study the cause of their migrations; and car movement data can be integrated with air pollution data or weather data to study impacts and so on. When it comes to integration of remote sensing data with trajectories, one would commonly rely on a conventional chain of GIS operations to match trajectory locations to grid values: download grid data, georeference, match each trajectory record to a corresponding image cell (or cells), perform overlay, extract cell values for a given location and time and compose values into a resulting table. Clearly, when dealing with large and dynamic spatio-temporal records this approach is unmanageable. In the paper we demonstrate an example of utilizing web services for trajectory and grid data integration and show how this integration service can be embedded into distributed geospatial components for manipulation and visualization of trajectory data.

Christine Plumejeaud*, Jérôme Gensel*, Hélène Mathian** and Claude Grasland**
* : Laboratoire d'Informatique de Grenoble (LIG), UMR 5217, 681 rue de la passerelle, BP. 72, 38402 St Martin d’Hères, France christine.plumejeaud@imag.fr jerome.gensel@imag.fr
** : Laboratoire Géographie-cités, UMR 8504, 13, rue du Four 75006 Paris mathian@parisgeo.cnrs.fr. claude.grasland@parisgeo.cnrs.fr

Nowadays, many data sources provide timestamped information for socio-economical indicators, that could highlight evolutions of territories over time. Every body would dream of a system that integrates the various data sources and provide a cartographic tool for spatio-temporal analysis. Such a system requires a data model that can handle the various problems linked to the integration of time into geographic information systems, such as the evolutivity of the geographic support, and the change of semantics of the statistical variables through the time. It also requires some new concepts for spatio-temporal analysis in order to show on maps the evolution of territories, relatively to a spatio-temporal context. In existing systems, the user can visualise data onto maps at each date of census, and even get an interpolated curve presenting the evolution of the data through time; but something is still lacking that could automate comparisons and bring observations into their spatio-temporal context. What is needed is ways to explore in order to visually compare the evolution of a given territory with territories where neighboorhood, hierarchical and genealogical relationships play an important role. This paper focuses on problems linked to spatio-temporal analysis. First, we present a data model handling such relationships between territories. Then, we give some ideas concerning the exploitation of such relationships for a spatio-temporal analysis based on maps visualisation. In addition, we present an overview of the produced maps and reports based on the model we have integrated into a prototype, named HyperTime, which results from the adaptation of a previous software designed for spatial analysis, called HyperAtlas

G. Andrienko, N. Andrienko, S. Bremm, T. Schreck, T. von Landesberger, P. Bak, D.Keim
G.Andrienko and N.Andrienko are with University of Bonn & Fraunhofer IAIS, Germany
S.Bremm, T.Schreck and T.von Landesberger are with Technische Universitt Darmstadt & Fraunhofer IGD, Germany
P.Bak and D.Keim are with niversity of Konstanz, Germany

Spatiotemporal data pose serious challenges to analysts in geographic and other domains. For a comprehensive analysis, the data need to be considered from two complementary perspectives: (1) as spatial distributions (situations) changing over time and (2) as profiles of local temporal variation distributed over space. We suggest a framework based on the ”Self-Organizing Map” method combined with a set of interactive visual tools supporting both analytic perspectives. In the first perspective, SOM is applied to the spatial situations at different time moments or intervals. In the other perspective, SOM is applied to the local temporal evolution profiles.
The framework has been validated on a large dataset with real data about forest fires in Italy that happened over 23 years, where expected spatiotemporal patterns have been successfully uncovered. We also describe the use of the framework for discovery of previously unknown patterns of different distribution of forest fires according to their causes.

L. Paolino, M. Romano, M. Sebillo, G. Vitiello
Dipartimento di Matematica e Informatica, Università di Salerno (Italy) {lpaolino,marrom, msebillo, gvitiello}@unisa.it

In case of emergency, visual analytic applications may be a successful means for quickly organizing the rescue activities. They allow decision-makers to immediately visualize the status of the crisis, plan the evacuation and address people towards vacancies in emergency centres. Although the effectiveness of such application is immediately clear, further support may be gained by allowing people to directly manage the emergency on site. In this sense, it seems to be particularly desiderable to provide interfaces which support visual analytic tasks on small and handheld devices without losing their communicative efficacy. In this paper, we describe a new visual techinque, named Framy, which allows users to visualize in a very intuitive way a large number of aggregators on very small devices and which is suitable for the management of this kind of emergencies. The Framy visualization approach is explicitly targeted to enhance geographic information visualization on small-sized displays. The rationale behind Framy is to display semi-transparent colored frames along the border of the device screen to provide information clues about different sectors of the on/off-screen space. For a given query, the color intensity of each frame portion is proportional to an aggregating value of the objects located in the corresponding map sector either inside or outside the screen. Thus, for instance, the frame may indicate both the distance and the direction of specific Point Of Interests (POIs) but it may also represent the amount of POIs located towards a specific direction or the sum of a given attribute values. Framy represents a summary of data located in a given direction by means of the frame portion color intensities. Moreover, the number of frame portions can be interactively increased so as to refine the query results, indicating, e.g., the exact direction where certain objects can be found. If several aggregators are displayed on the map, nested frames may be visualized along the borders, each one corresponding to a different aggregator, with a different colour. Once the user performs the spatial requests, results are visualized on the map by applying Framy which helps users to analyze results. The user may choose to divide the map into any number n. Starting from the centre of the screen a fictive circle is drawn and the map is divided into n sectors of equal width (360°/n ). Fig. 1 gives an idea of this subdivision using n = 8. Given a sector Seci, we use the notation Ui – invisible- to represent the parts within Seci which lie outside the screen, whereas Vi – visible- represents the inside screen parts. The idea underlying this approach is to inscribe a semi-transparent area (frame) inside the screen. This area is partitioned in C1,…, Cn portions, each identified by the intersection between the frame and the Vi parts. Once the user poses a query, the color intensity of each Ci, which corresponds to its saturation index in the Hue-Saturation-Value (HSV) model, is modified on the basis of the value which aggregates a property of objects in either Ui or Vi or Ui ? Vi, such as sum and count, or calculates the distance between the map focus and a POI in one of this set. Namely, the color intensity may be expressed as
intensity(Ci) = s(g((Si)) for each i in {1, …, n} (1)
where s is a monotonic function, g is a function which aggregates starting from a set of spatial data and Si is one of either Ui or Vi or Ui ? Vi (from now on aggregator). Starting from the HSV model, we have specified how its components may be calculated in order to determine the color value of each Ci. In particular, we set the Value parameter to 100, thus implying that the highest color brightness is always associated with Ci. The Hue component corresponds to the color assigned to the whole frame. Finally, the saturation index corresponds to the color intensity and it is determined on the basis of the type of the function g in (1).

Ilya Boyandin (University of Fribourg), Enrico Bertini (University of Konstanz), Denis Lalanne (University of Fribourg)

This paper presents JFlowMap, a graphical tool offering various visualization techniques for producing and analyzing flow maps, one of the most often used visualizations of migration flows. We show how these techniques can be used in combination to create task-oriented views, i.e. views that help to solve specific user tasks. Further, we demonstrate on the UNHCR Refugee Dataset how our tool can be applied to the analysis of temporal changes in migration flows, and finally, discuss the possible directions of future work.

Diansheng Guo, Shufan Liu, Hai Jin
Department of Geography, University of South Carolina (709 Bull Street, Columbia, SC 29208)

It is difficult to visualize and extract meaningful patterns from massive trajectory data. One of the main challenges is to characterize, compare, and generalize trajectories to find general patterns and trends across space and time. Existing methods often use a vector-based approach, which compares and groups trajectories (or sub-trajectories) based on segment locations (distances), time differences, speeds, and angles. Another challenge is to generalize individual locations into regions of interest. Existing methods normally use a density or distace-based approach to aggregate locations or grid cells into larger areas, which cannot adapt to the spatial variation of pattern densities. This research proposes a graph-based approach to address the above two challenges, which treats trajectory data as a complex network. Within the context of urban transportation and vehicle movements, the research incorportes road networks to establish topological relationships among trajectories and locations. A graph-based regionalization and community detection method is then used to find hierarhical spatial structures in the movements. Finally, trajectories are visualized and explored in both space (at various region levels) and time (at different time intervals).