Next Week a the AAG Annual Meeting

Next week, I will attend the 2008 Association of American Geographers Annual Meeting in Boston where I will present my work The co-evolution of taxi drivers and their in-car navigation systems at the session Situating Sat Nav 2 (from 4:20 PM – 6:00 PM). Description of the session:

Sat Nav offers technologically sophisticated spatial data models of the world, but the technology quickly sinks into taken-for-granted everyday driving practices, such that its social and political significance is hard to assess. The gadgets themselves take space on the dashboard and windscreens, but also make new senses of space for the driver, well beyond the car. The session will present a range of theoretically informed analyses questioning the social effects, cultural meanings and political economy of in-car satellite navigation.

Other sessions I plan (or wish) to attend are:

Tuesday
“New” Geographies of Mobility and Accessibility: Theory, Modelling, and Policy Implications (from 12:00 PM – 1:40 PM)
During the 1950s, Ullman and Mayer prepared an initial sketch of the areas of knowledge specialization emerging from the intersection of geography and transportation. Their work provided a framework for the development of Transportation Geography. Among the various themes they identified, there was an emphasis on the study of systems, flows, and interactions. Mobility, flows, and the production of capital were physical processes involving place-based production of goods and services, and the physical movement of commodities and people through time and space. Accessibility was a product of location among origins and destinations of those commodities and people. Today our conceptualization continues to evolve in the face of wireless and wired technologies. We are at times both the producers and consumers of our own wares (Toffler, 1980), and increasingly engage in the use of information and communication technology (ICT) to perform obligatory and discretionary activities, and to consolidate and extend our social networks. In the face of the sort of spatial deconstruction offered by what Sheller and Urry (2006) and others have called, “the new mobility”, Transportation Geographers and those in many other disciplines face new challenges and research opportunities as they attempt to come to grips with the relationship between mobility, accessibility, space, and place in the information age. This session will explore recent theoretical, qualitative, empirical, and policy-based discourse and practice surrounding emerging geographical perspectives regarding relationships between technology, mobility, accessibility and daily life.

Wednesday
Urban Geography: Urban Processes and Models (from 8:00 AM – 9:40 AM)

Spatial Data Analysis, Visualization, and Modeling (from 1:00 PM – 2:40 PM)

Time Geography: Emerging Theoretical Developments, Implementations, and Applications (from 1:00 PM – 2:40 PM)
Originally designed to investigate various constraints of human activities in time and space, the time-geographic framework provides an integrated space-time environment to effectively and efficiently investigate the spatio-temporal characteristics of human activities and their interactions. There have been revived research interests in time geography in recent years. These research efforts include extending the time-geographic framework to accommodate the emerging hybrid environment of physical and virtual spaces, providing computational models and representations of the framework, developing GIS designs to implement the framework, and applying the framework to facilitate studies such as travel behaviors, activity patterns, accessibility assessment, urban structure, animal ecology, etc. This session will provide researchers a forum to share experiences and exchange ideas on recent theoretical developments, implementations, and applications of time geography.

Geographies of Play III: Embodied, emotional, sensory geographies of play (from 1:00 PM – 2:40 PM)
Christopher Harker (2005: 59) reminds us “Playing is not (just) kids stuff. Playing is something we all do, albeit to different extents and degrees, and this is something that needs a great deal more investigation”. These sessions respond to Harker’s appeal for more critical attention to be given to the study of the geographies of play. The sessions include papers from a wide range of disciplines and perspectives, to encompass child, adult and intergenerational experiences of play. The papers explore innovative ways of studying the geographies of play and embrace a wide range of diverse theoretical and methodological approaches.

Visualization (from 3:10 PM – 4:50 PM)

Urban Tourism (from 3:10 PM – 4:50 PM)

Thursday
Urbanism and Urban Planning (from 8:00 AM – 9:40 AM)

Agent Based Modeling, Simulation (from 8:00 AM – 9:40 AM)

Spatial Analysis and Modeling: Transport and Spatial Analysis (from 10:10 AM – 11:50 AM)

Cyberinfrastructure-Data and Knowledge Representation (from 10:10 AM – 11:50 AM)
The flourishing developments of shared geographic data, information, knowledge and computing resources have produced many products to facilitate the easy use of geographic resources. For example: 1) Google Earth and Microsoft Virtual Earth have changed how we explore geographic extent; 2) OGC developed multiple web services to facilitate communication among GIS components that are widely used in assembling services, such as spatial web portals; 3) Geographically distributed sensor webs have opened up the possibilities for real-time control of complex systems such as urban traffic; 4) Knowledge representation systems enable the enterprise to accumulate knowledge and make smart decisions. These evolutions adopt cyberinfrastructure to facilitate geographic research, development, and education.

Applied geostatistics (from 1:00 PM – 2:40 PM)
This session will provide an overview of the state-of-the-art in the application of geostatistics to a wide variety of disciplines.

Subversive cartographies (from 1:00 PM – 2:40 PM)
Subversive cartographies is a series of sessions jointly organised with the Maps in Society Commission of the International Cartographic Association. This first session brings together papers emphasizing the role of the aesthetic in the construction of alternative and artistic mappings. Common themes are the relations between artistic practice and mapping, narrative and (e)motional cartographies, and the politics of design.

Friday
Internet Mapping and Mash Ups (from 10:10 AM – 11:50 AM)

A Conversation with Noam Chomsky (from 2:30 PM – 4:10 PM)
Saturday
Visualization, Cartography, and Cognition (from 8:00 AM – 9:40 AM)
Build it, Mapt it, Web it (10:10 AM – 11:50 AM)

Visualization: Viewing Data in New Ways (from 10:10 AM – 11:50 AM)

Program of the Round Table on Real-Time Cities

So, ok the program for Monday’s Round Table on Real-Time Cities is now set. The informal aspect of the event should make it easy to rearrange it on the fly if necessary.

I will start of by introducing and defining the subject; Mentioning that, cities are by definition real-time, but the deployment of geo-information, mobile, wireless and sensor technologies allow to reveal the global, emerging aspects that can be reacted upon. In other words, a “real-time city” is a city in which system conditions can be monitored and reacted to instantaneously (Townsend, 2000); Making real the narrative of Archigram that suggested that the way the street feels may soon be defined by what cannot be seen with the naked eye “When it’s raining on Oxford Street, the buildings are no more important than the rain”. The visions behind “real-time cities” often refer to “pulsing cloud of data, instantaneous information, seamlessness integration, empowerment of the citizens, enhancement of our perception, reveal the city as we experience it, patterns of behavior, observe and improve“. For instance here and there I could find sentences such as: “Seamless integration of real-time information about events, resources, and personal experience within physical spaces” or “Strengthen our perception of the built environment as a place for social inclusion and collaboration” or “Before transport planning was about predict and accommodate and now it becomes more observe and improve“. This round table takes the opportunity to gather researchers in urban planning, geographic information systems, architecture, computer science, social sciences, and interaction design and share our perspectives on this new object of research. In 3 hours there will probably be no much more time than to break the ice and raise an awareness of the multiple issues inherent to the design, deployment and integration of real-time information systems in cities.

The session will be split in 2 main parts. At a first step, informal presentations and discussion around 3 topics:

Topic 1: New resources to describe cities (talks: Raj Singh, Paul Torrens)
Topic 2: The city as a platform for innovation (talks: Jonathan Raper, Georg Gartner)
Topic 3: Implications of the deployment of ubicomp technologies on the reconfiguration of cities (talks: Adam Greenfield, Carlo Ratti)

Then, in a second part, as long as time allows, discussion on three stakeholder perspectives

Citizens: What is a good real-time city?

Research: How can we define this research object? What drives us?
Practitioners: What are the expertise to design/manage a real-time city? What is the process (action vs. reaction)?

The round table will be directly followed by an MIT open lecture by Adam Greenfield entitled “The City Is Here For You To Use“.

Relation to my thesis: I guess at no other places than MIT, I would have been able to gather such a top set of people to discuss themes at the core of my thesis: digital traces, revealing the invisible, urban modeling and simulations, user-centered datasets and location-based services.

The Repast Simulation/Modelling System for Geospatial Simulation

In The Repast Simulation/Modelling System for. Geospatial Simulation, CASA’s Andrew Crooks (blog) introduces Repast and its ability to incorporate geospatial data to build an environment and to create agents to explore different types of phenomena. He shortly introduces some geosptatial simulation constructed from minimalist academic models based upon ideal assumptions, to large scale decision support systems based upon real-world data. These include:

• disaster management such as a Sarin attack in Manhattan, the spread of infectious diseases and a food poisoning outbreak
• pedestrian modelling both in retail and emergency evacuation
• urban dynamics (see Barros, J. (2004), Urban Growth in Latin American Cities: exploring urban dynamics through agent-based simulation), segregation, residential and business location
• models in which mobile agents travel and interact on rugged terrain or on network landscapes at any scale from rooms within buildings to urban neighbourhoods to large geographic networks of cities. (see GeoGraph)
• study land use change
• geographic conflict research(see GROWLab)

Relation to my thesis: Exploring the possible use of “census data” generated by digital footprints (user-generated data, cellphone data) to model urban tourism (some profiles of tourists) dynamism both in time and space. It would be one way to validate the data (compare the simulation with survey data) and test different scenarios over several cities, thus aiding planning tourism. Appetizer before reading Cities and Complexity: Understanding Cities with Cellular Automata, Agent-Based Models, and Fractals.

Recreation Behavior Modeling and Simulation

In the processing of building a coherent story around my research endeavors and considering potential outcome, I have returned to exploring how agent-based modeling techniques can help grapple with the validation and significance of user-genereated content in the realm of urban/mobility/tourist research. Current tourism simulation and modeling (see for instance TourSim) works mainly rely on specific surveys to build and evaluation the simulation. In addition, the data collected describe tourist behaviour such as spending habits, and psychological motivations for tourism. These sparse information make it hard to reflect the complexities of tourist behavior and build effective and efficient decision support tools to assess planning decisions. What is required for recreation planning, is verification of how tourists act spatially at recreation sites. However key variables such as the speed of tourist travel, wayfinding decisions, crowd avoidance, and other spatial behaviour, are not yet well understood to model the tourist visiting a city. One of my hypothesis is that digital footprints such as user-generated content can help develop agent-based models and simulations of tourist flows and movements (in that case through photography).

Similar to transport research, some tourist research collect quantitative data of tourist activity such walking and photography. In Building better agents: Geo-temporal tracking and analysis of tourist behavior the authors use quantitative data captured by sensors to build agent-based models of tourist behaviors. Their simulation provide one way for managers to accurately predict future impacts, and their spatial patterns of the develop of certain tourist areas. They analyze:

• detailed visitor counts
• average trip durations
• tourist behavior
• spatial patterns of movement

in order to reveal some group and individual behaviors:
Crowding: Determing through correlation whether people were spending less time, for example, on the viewing platform, during more crowded times of the day.
Graphing: Provide detailed information about times and sequences of travel for individuals and groups
Travel time: Time frequency distribution to be analyzed. Correlation between time spent in various area of the study site
Travel sequence: Tourist behavior can be devided into distinguishable groups based on movement sequences

However, the overall validity of the simulations remains uncertain without detailed calibration data. As described in Understanding of tourist dynamics from explicitly disclosed location information, the flickr dataset can provide more coarse grained quantitative observations of similar phenomenon. However, user-generated data can surpass the scalability and time constraints of surveys and sensor-based approaches. My current believe is that the availability of data over the world’s most photographed cities can allow me to validate tourist models build from user-generated content. Building such a model and validate it with simulation over several cities might be one nice outcome.

Next steps in that direction, Michael Batty wrote a book on Cities and Complexity: Understanding Cities with Cellular Automata, Agent-Based Models, and Fractals. Repast Simphony 1.0 has recently been released which includes a point and click interface for model development and full GIS support.

Life in the Real-time City: Mobile Telephones and Urban Metabolism

Townsend, A. M. (2000). Life in the real-time city: mobile telephones and urban metabolism. Journal of Urban Technology, 7(2):85–104.
Back in 2000, Anthony Townsend wrote Life in the real-time city: mobile telephones and urban metabolism, an article that argues that new mobile communication systems are fundamentally rewriting the spatial and temporal constraints of all manners of human communications. As signs of this radical change, accessibility becomes more important than mobility and mobile phones increasingly add an element of uncertainty about physical location to our urban interactions. For instance, as many as one-fifth of cell phones users lie about their location when talking on a mobile phone. “For urban planning, it might mean that the city will change far faster than the ability to understand it from a centralized perspective, let alone formulate plans and policies that will have the desired outcomes”.

As decision-making and management of everyday life is increasingly decentralized, the complexity of these systems become greater and therefore less predictable. In parallel, this decentralization creates myriad new interactions and potential interactions between individuals that is dramatically speeding the metabolism of urban systems, increasing capacity and efficiency.The “real-time city” in which system conditions can be monitored and reacted to instantaneously, has arrived. [...] Real-time systems are defined by an ability to constantly monitor environmental conditions vital to the operation of the system.

In fact without efforts to develop new knowledge and tools for understanding the implications of these new technologies, city planners run the risk of losing touch with the reality of city streets.. Townsend takes an urbanist’s perspective on the application of new communication technologies within cities by their inhabitants (i.e. how do they reshape basic aspects of urban life). However contrary to traditional urban planning, which often assigns agency to a city as a unit (e.g. the city is busy, the city in unfriendly), there are tools for understanding complex systems like cities as consequences of many interactions of individuals. Yet, these tools must go beyond the classical approaches taken in urban planning “the widespread bit-by-bit reconstruction of cities is going largely unnoticed by planners accustomed to visualizing cities through aerial photographs“. In consequence, individuals must become the unit of analysis instead of the institution, neighborhood, city or region. These types of new insights can be gained from interpretive methods such as ethnography (e.g. Kevin Lynch’s The Image of the City) or a psychoanalysts approach. Then the significance of individual-level technological interventions on larger-scale social systems such as cities could be simulated through agent-based modeling.

Also blogged by Nicolas in Increasing pace of interactions in our cities.

Relation to my thesis: This text refers to some pieces of my works. First, similarly to Antoine Picon’s suggestions last week, Townsend stresses the focus on individual interactions (micro events) that make the city. There is also a reference to some sort of glocalization of the city generated by the telephone and mobile phones (decentralizations in urban sprawl and intensification of the center). Second, there is a practical discussion on how taxi driver’s archaic profession was transformed by mobile phones. “The mobile phone permits dynamic reallocation of the taxi system’s resources, resulting in less wasted time searching for fares“. Something that I can argue with my taxi driver study and the importance of satnav not only to improve the efficiency, but also to decrease the stress (improve the quality of life). Third, there is a reference that mobile technologies add uncertainty to our urban interactions (CatchBob!). Finally, this text revives agent-based modeling as a potential output of my thesis.

Analyse et Modélisation des Activités Coopérative Situées

Salembier, P. & Pavard, B. (2004) Analyse et modélisation des activités coopératives situées. Evolutions d’un questionnement et apports à la conception, @CTIVITES, n°1, Vol.

Cet article retrace les évolutions d’un programme de recherche centré sur l’analyse, la modélisation et l’instrumentation des activités coopératives. Dans son fil conducteur, il introduit plusieurs points (explication des arrière-plan théoriques, méthodes d’appréhension des objects théoriques ciblés, nature des modèles produits, rapport à la conception des situations) important à considérer lors de la réflexion générale sur l’édute des activités professionnelles, collectives et coopératives. Les auteurs pointent les avantages et les limites de différent orientations théoriques et méthodologiques ayant balisé leur travail.

Mentions sont faites des différentes conceptions de la cooperation homme-machines. Une première orientation illusior et l’approche mimétique dans laquelle la reproduction des caractéristiques des situations de coopération homme-homme se heurte à des limitations technologiques, mais également à des problèmes de font qui ont trait à la non reconnaissance de la profonde asymétrie des partenaires . Ceci est en fait le thème de “Plans and Situated Actions” de Lucy Schuman. Une alternative est “l’allocation de fonction“, c’est-à-dire procéder à la répartition des tâches entre opérateur et système (MABA-MABA: “Men are Better At – Machines Are Better At”). Cette approche a donné lieu à des recherches autour de la notion de “système cognitif joint” où l’artifact informatique tient un rôle d’”outile congnitif“. Une approche pragmatique est l’ingénieurie cognitive qui fait appelle à la “Triade du système cognitif” (Woods, 1998) avec ces trois facteurs: le monde, l’agent (humaine, machine, système “hybride”) qui opère sur les monde et les instances de méditisation du réel. L’idée sous-jacente est que la concéption de supports améliorant l’éfficacité de la réalisation de la tâche passe nécessairement par la compréhension des interactions entre ces trois éléments. Les exigences et contraintes sont pris en compte lors de la construction d’une description cognitive autonome de l’environnement (facteurs augmantant la complexités cognitive de la tâche tels que la nature dynamique de l’environnement, nombre important de parties interconnectées, incertitude des données, ..). Tout déséquilibre de la Triade risque de se traduire par l’émergence d’une situation dangereuse ou non désirée doit alors faire l’objet d’une réponse adaptée. Les méthodes utilisées pour mener à bien le processus de mise en relation systématique entre exigences de la tâche, contraintes imposées par l’environnement et ressources cognitives mobilisables par l’agent, ont été synthétisées sous le terme de “Cognitive Task Analysis“. Le point important est que les situations sont spécifiées en terme cognitifs et non pas comme en ingénieurie classique dans les termes du dispositif technique. Le rôle de cette analyse de l’activité permet d’identifier les bottlenecks dans le système, c’est-à-dire les limitations cognitives dans le couplage opérateur-outil et les contraintes de l’environnement pointées par les opérateur.

Référentiel contextuel
Les auteurs se sont alors centrés sur les mécanismmes informels de la coopération homme-homme, plus précisement sur la construction et l’actualisation dans le cours de l’activité d’un référentiel contextuel partagé qui constitue une des conditions du déploiement efficace de l’activité collective dans un environnement complexe distribué et la régulation par ce collectif de facteurs tels que les variations dans la charge de travail et la fiablilité globale du système socio-technique.

Rôle des artefacts
Les premières analyses sont restés très superficielles quant aux propritétés physiques des objects utilisés par les opérateur. Les objet étaient essentiellement considérés dans leur dimension informationnelles (artefacts cognitifs) et non dans leur dimension manipulable (constituvité matérielle des activités cognitives)

Modélisation et simulation
La construcction de modèles n’est pas jugée nécessaire voire utile par les ethnométhodologue. Les auteurs ont tout d’abord utilisé la modélisation pour décrire et non simuler. Puis ils ont recouru à la simulation:

L’idée est ici plutôt de jouer sur un ensemble de variables exogènes d’environnement pour explorer un changement de situation (introduction d’un nouvel outil, modifications des formes d’organisation du collectif, …) et évaluer son impact sur des variables endogènes jugées pertinentes -par exemple une appréciation quantitative du contexte partagé- (Zorola-Villarreal, Pavard, & Bastide, 1995 ; Salembier, Kahn, Zorola-Villarreal, & Zouinar, 1997). La simulation fonctionne comme un moyen pour « ouvrir l’espace de conception » et nourrir l’interaction (discussion et négociation) entre les acteurs engagés dans la conception autour de différentes alternatives possibles.

Les auteurs vont alors passés des outils logiques classiquement utilisés en sciences cognitive aux théories de la complexité et aux systèmes dynamiques non linéaires. La raison qui movitve ce choix tient à l’acceptation du fait que la dynamique des processus qui supportent la coopération sont la plupart du temps impossilbe à prévoir car non déterministes du fait notamment de leur caractère distribué et de leur sensibilité aux vations environnementales. (Pavard & Dugdale, 2000).

La démarche mise en oeuvre se résume à: analyse de la tâche, analyse de l’activité, simulation papier-crayon, simulation partielle de la situation, mise en situation recréée, simulation informatique.

Limites
Les limites de l’approche sur la modélisation multi-agent est se concentrer essentiellement sur l’émergence de propriétés globales sans pouvoir les articuler avec les comportements locaux des agents ; le spectre d’un collectivisme méthodologique radical n’est pas loin… De plus, la dimension « située » des acteurs, leurs connaissances culturelles, leur intelligence contextuelle, … sont des dimensions difficiles sinon impossibles à représenter de façon synthétique dans les agents distribués (Dugdale & Pavard, 2002).

Dans la suite de l’étude, les outils utilisés sont essentiellement des environnements de simulation multi-agents qui permettent de recréer la dynamique du fonctionnement du collectif en réponse à des modifications de l’environnement externe (survenue d’événements particuliers) ou interne (modification de la structure du collectif d’agents).

Relation to my thesis: Je m’étais éloingné des aspects collectifs et coopératif de l’utilisation d’applications géolocalisées et de leur étude en situation “naturelle” ainsi que de l’utilisation d’ABM. Dans leurs études, Salembier et Pavard utilisent la simulation comme un moyen pour « ouvrir l’espace de conception » et nourrir l’interaction (discussion et négociation) entre les acteurs engagés dans la conception autour de différentes alternatives possibles. Ils ont fait appel à des champs disciplinaires connexes (théorie des actes de langage, théorie des systèmes complexes, éthologie, théorie des systèmes multi-agents, interactionnisme, …).

Design for Emergent Behaviour with Agents

Related to my thoughts on Wireless Mobile Agents and the Internet of Things, O’Reilly’s Paul Browne suggests that a agent computing might be a paradigm to replace OO programming. He mentions Cougaar for creating and managing agents.

Relation to my thesis: Ubicomp infrastructures are centrally controlled. Some could be designed for the emergence of behaviours as agent programming naturally suggests.

Agent Modelling and Simulation Presentation

Yesterday, as part of a doctoral school course, I gave a short presentation on Agent-Based Modelling and Simulation. The talk was mainly focused on a brief theoretical background (differences ABM and conventional models, the methodological approach) example of ABM in social sciences (Schilling). I developed and simulated a basic model I called “coopetitive famers” (coopetition coming from the merge of cooperation and competition) as a heuristic approach to competitive environments with cooperative and individualistic farmers. It was a way to show patterns emerging from basic interaction rules (agent mainly interacting with the environment).

I was advised to have a look at “The Evolution of Cooperation” which discusses how cooperation can emerge in a world of self-seeking egoists when there is no central authority to police their actions.

Relation to my thesis: I worked on this to grasp the possibility to introduce ABM into my thesis as a methodology to get insights on users and groups behaviors in a mobile and ubiquitous environment. For example try to model their interaction rules in cases of zero/bad/avg/good positioning accuracy or latency. As a next step, I need to investigate on the way to integrated GIS into ABM.

Pedestrian Modelling and Movement Analysis

Related to my current interest in modeling users of mobile and ubiquitous environments, Intelligent Space Partnership is working in urban areas on pedestrian movement analysis and modelling. Their developments lead to better understanding of pedestrian movement, shopping patterns and advise on the optimisation of public space design.

Pedestrian-oriented land uses in central Leeds (left), street network accessibility analysis in Washington D.C.

Why Agents? On the Varied Motivations for Agent Computing in the Social Sciences

Robert Axtell. 2000. Why Agents? On the varied motivations for agent computing in the social sciences.

This paper (mainly focused in social sciences and political economy in particular) argues the existence of three distinct uses of agent-based computational models:

• When numerical realizations are relevant, agents can perform a variant of classical simulation
• When a model is incompletely solved mathematically, then agent-based model can be a useful tool of analysis, a complement to mathematics. It is generally possible to build agent-based computational models in order to gain insight into the functioning of the model.
• There are cases in which mathematical models are either apparently intracable or provably insoluble. Agent-based computing is perhaps the only technique available for systematic analysis, a substitute for formal mathematical analysis.

Strengths and Weaknesses
A very common motivation for ABM is dissatisfaction with rational agents. In most social processes either physical space or social network matter. These are difficult to account for mathematically except in hightly stylize ways. However, in ABM, it is usually quite easy to have the agent interactions mediated by space or network or both. Spatial networks are quite naturally represented in agent-based computatinal models. A physical location can be part of an agen’ts internal states. Likewise, its position in a social network can be easily represented internally. In ABM the only way to prove a sufficiency theorem is to go through multiple runs, systematically varying initial conditions or parameters in order to assess the robustness of results.