Skip to main content
Location-aware data collection technologies provide new insights about location choices. Only a few dynamic models of location choice exist in scientific literature. To our knowledge, none of them correct for serial correlation. In this... more
Location-aware data collection technologies provide new insights about location choices. Only a few dynamic models of location choice exist in scientific literature. To our knowledge, none of them correct for serial correlation. In this paper, we model choice of catering locations on a campus using WiFi traces. We use the Wooldridge (2005) correction method that deals with the initial values problem and related endogeneity bias in estimation. Cross-validation, price elasticity and simulation of a scenario predicting the opening of a new catering location are presented. Predicted market shares of the new catering location correspond to point-of-sale data of the first week of opening.
Location-aware data collection technologies provide new insights about location choices. Only a few dynamic models of location choice exist in scientific literature. To our knowledge, none of them correct for serial correlation. In this... more
Location-aware data collection technologies provide new insights about location choices. Only a few dynamic models of location choice exist in scientific literature. To our knowledge, none of them correct for serial correlation. In this paper, we model choice of catering locations on a campus using WiFi traces. We use the Wooldridge (2005) correction method that deals with the initial values problem and related endogeneity bias in estimation. Cross-validation, price elasticity and simulation of a scenario predicting the opening of a new catering location are presented. Predicted market shares of the new catering location correspond to point-of-sale data of the first week of opening.
Research Interests:
To tackle the challenges of the 21st century, future scientists and engineers have to understand the interplay between societal challenges and technical solutions as early as possible in their education. They also have to develop the... more
To tackle the challenges of the 21st century, future scientists and engineers have to understand the interplay between societal challenges and technical solutions as early as possible in their education. They also have to develop the communication and the teamwork skills required to be effective professionals. To address this issue, the Ecole Polytechnique Fédérale de Lausanne (EPFL) introduced a new Global Issues program to all 1800 first year engineering students. In this paper, we present this novel program and reflect on our experience. Our results suggest that student who showed positive attitude towards teamwork, benefited the most from the course and increase their perspectives on societal issues as measured by their moral reasoning after the course.
Research Interests:
This thesis develops models of activity and destination choices in pedestrian facilities from WiFi traces. We adapt the activity-based travel demand analysis of urban mobility to pedestrians and to digital footprints. We are interested in... more
This thesis develops models of activity and destination choices in pedestrian facilities from WiFi traces. We adapt the activity-based travel demand analysis of urban mobility to pedestrians and to digital footprints. We are interested in understanding the sequence of activities and destinations of a pedestrian using discrete choice models and localization data from communication antennas. Activity and destination choice models are needed by pedestrian facilities for decision aid when building new infrastructure, modifying existing infrastructures, or locating points of interest. Understanding demand for activities is particularly important when facing an increasing number of visitors or when developing new activities, such as shopping or catering. Data from existing sensors, such as WiFi access points, are cheap and cover entire facilities, but are imprecise and lack semantics to describe moving, stopping, destinations or activities carried out at destinations. Thus, understanding pedestrian behavior first requires to observe the actual behavior and detect stops at destinations, and second to model the behavior. Part I of this thesis focuses on activity-episode sequence detection. We develop a Bayesian approach to merge raw localization data with other data sources in order to take into account the imprecision and describe activity-episode sequences. This approach generates several activity-episode sequences for a single individual. Each activity-episode sequence is associated with a probability of being the true sequence. The prior represents the attractivity of the different points of interest surrounding the measurement and allows the use of a priori information from other sources of data (register data, point-of-sale data, counting sensors, etc.). Part II proposes models for activity and destination choices. The joint choice of activity type and activity timing is modeled by seeing a sequence of activity episodes as a path in an activity network. Time is considered as discrete. Unlike traditional models, our model is not tour-based, starting and ending at the home location, since the daily ``home''activity is meaningless in our context. The choice set contains all combinations of activity types and time intervals. The number of different paths is thus very large (increasing with time resolution and disaggregation of types of activities). Inspired by route choice models, we use a Metropolis-Hastings algorithm for the sampling of paths to generate the choice set. An importance sampling correction of the utility allows the estimation of unbiased model parameters without enumerating the full choice set. While the activity path model describes the choice of an activity type in time, the location where the activity is performed is modeled with a destination choice model conditional on the activity type. Our approach accounts for the panel nature of the data and deals with serial correlation between error terms. Using real WiFi data collected on the EPFL campus, we detect pedestrian activity-episode sequences, estimate an activity path choice model and develop a destination choice model for a specific activity type: eating. Knowing that the individual has decided to eat, which restaurant does she choose? This conditional destination choice model includes in its utility the cost of menus, available types of foods and drinks, distance from a previous activity episode, socioeconomic characteristics and habits.
Research Interests:
While moving from diary survey to location-aware technologies, recent data collection techniques provide new insights about location choices. Only few dynamic models of location choice exist in the literature, and none of them to our... more
While moving from diary survey to location-aware technologies, recent data collection techniques provide new insights about location choices. Only few dynamic models of location choice exist in the literature, and none of them to our knowledge correct for serial correlation. In this paper, we apply a method proposed by Wooldridge (2005) to deal with the initial values problem on the choice of catering locations on a campus using WiFi traces. Cross-validation, price elasticity and simulation of a scenario predicting the opening of a new catering location are presented. Predicted market shares of the new catering location correspond to point-of-sale data of the first week of opening.
Download (.pdf)
We propose a model for the choice of an activity pattern. Models of activity participation patterns allow to assess the impact of demand management strategies on activity and destination choices. In particular, we focus on choice set... more
We propose a model for the choice of an activity pattern. Models of activity participation patterns allow to assess the impact of demand management strategies on activity and destination choices. In particular, we focus on choice set generation of activity patterns using recent developments in route choice modeling. Spatial choices deal with large choice sets. We develop a framework for choice set generation based on path choice. The activity-episode sequence is modeled as a path in an activity network defining the activity type, duration and time of day. The large dimensionality of the choice set is managed through an importance sampling based on Metropolis-Hastings algorithm. Our model can be used to forecast demand at the urban scale and also in pedestrian facilities, such as transport hubs or mass gathering. Validation of the approach is performed on synthetic data and a case study using WiFi traces on a campus is presented.
Research Interests:
In this paper, we propose a methodology to use the communication network infrastructure, in particular WiFi traces, to detect the sequence of activity episodes visited by pedestrians. Due to the poor quality of WiFi localization, a... more
In this paper, we propose a methodology to use the communication network infrastructure, in particular WiFi traces, to detect the sequence of activity episodes visited by pedestrians. Due to the poor quality of WiFi localization, a probabilistic method is proposed that infers activity-episode locations based on WiFi traces and calculates the likelihood of observing these traces in the pedestrian network, taking into account prior knowledge. The output of the method consists of candidates of activity-episodes sequences associated with the likelihood to be the true one. The methodology is validated on traces generated by a known sequence of activities, while the performance being evaluated on a set of anonymous users. Results show that it is possible to predict the number of episodes and the activity-episodes locations and durations, by merging information about the activity locations on the map, WiFi measurements and prior information about schedules and the attractivity in pedestrian infrastructure. The ambiguity of each activity episode in the sequence is explicitly measured.
Gathering data about pedestrian origin, destination and route is difficult, particularly indoor and on a large scale. These data are important for route choice modeling, description of congestion, and flow estimation. Most data collection... more
Gathering data about pedestrian origin, destination and route is difficult, particularly indoor and on a large scale. These data are important for route choice modeling, description of congestion, and flow estimation. Most data collection techniques are device-centric. In this paper, we focus on the communication network infrastructure and propose to use WiFi traces to generate pedestrian destinations. Due to the poor quality of WiFi localization, a probabilistic method is proposed that infers visited destinations based on WiFi traces and calculates the likelihood of observing these traces in the pedestrian network, taking into account prior knowledge. The output of the method consists in generating several candidate lists of destinations, and assigning the probability of each list being the true one. Results show that it is possible to predict the number of destinations, the time spent at it and the localization of it, discriminating intermediary signals from signals generated at destination.
Cette étude se situe dans le cadre d’une recherche multidisciplinaire sur les pratiques modales en couronne des agglomérations mandatée par CarPostal avec pour objectif principal de mieux comprendre les pratiques de mobilité des usagers... more
Cette étude se situe dans le cadre d’une recherche multidisciplinaire sur les pratiques modales en couronne des agglomérations mandatée par CarPostal avec pour objectif principal de mieux comprendre les pratiques de mobilité des usagers actuels et potentiels de CarPostal. Les résultats de cette ample recherche ont mené l’équipe à s’interroger sur le lien entre l’offre et la demande et les effets concrets des augmentations de fréquences sur la fréquentation des lignes. Suffit-il d’augmenter la fréquence pour que la fréquentation suive ? Quels sont les facteurs ou les conditions favorables à une augmentation de la fréquentation ? Dans quels contextes cette tendance se confirme ou s’infirme ? Comment les régions font-elles face aux contraintes et aux défis liés à la planification de l’offre ? Le présent rapport permet de répondre à ces interrogations.
Download (.pdf)
This paper develops a framework for understanding pedestrian mobility pattern from WiFi traces and other data sources. It can be used to forecast demand for pedestrian facilities such as railway stations, music festivals, campus,... more
This paper develops a framework for understanding pedestrian mobility pattern from WiFi traces and other data sources. It can be used to forecast demand for pedestrian facilities such as railway stations, music festivals, campus, airports, supermarkets or even pedestrian area in city centers. Scenarios regarding the walkable infrastructure and connectors, the scheduling (trains in stations, classes on campus, concerts in festivals) or the proposed services in the facility may then be evaluated. It is inspired by activity-based approach. We assume that pedestrian demand is driven by a willingness to perform activities. Activity scheduling decision is explicitly taken into account. Activity-based approach for urban areas is adapted for pedestrian facilities, with similarities (scheduling behavior) and differences (no “home” in pedestrian facilities, thus no tours). This is a first attempt to define a integrated system of choice models in the context of pedestrian facilities.
Download (.pdf)
Ce rapport présente les résultats d’une enquête quantitative menée dans le cadre d’un projet de recherche dans le domaine de la mobilité combinée, baptisé "Optima" et confié par CarPostal au Centre de Transport (TraCE) de l’Ecole... more
Ce rapport présente les résultats d’une enquête quantitative menée dans le cadre d’un projet de recherche dans le domaine de la mobilité combinée, baptisé "Optima" et confié par CarPostal au Centre de Transport (TraCE) de l’Ecole Polytechnique Fédérale de Lausanne. Ce projet de recherche vise à décrire et comprendre les pratiques de mobilité des personnes vivant dans la périphérie d'agglomérations afin d'évaluer la demande pour des solutions innovantes en matière de transport public. Cette recherche s'appuie sur les résultats d'une précédente enquête, dite de préférences révélées, qui a permis de définir différents services, à la fois en termes d'information et de mobilité, susceptibles d’attirer et de fidéliser de nouveaux clients de CarPostal dans les régions étudiées. Un questionnaire de préférences déclarées a été envoyé aux répondants de la première enquête dans lequel ces services leur étaient proposés en remplacement de leur mode de transport habituel pour une boucle de déplacement qu’ils avaient eux-mêmes décrite dans l’enquête précédente, afin d'évaluer la demande pour ces nouveaux services. L’analyse des réponses a permis de mettre en évidence les préférences et les choix de la population concernée. Parmi les principaux résultats issus de cette étude, on observe une demande solide pour l'obtention d'information en temps réel, un net intérêt pour un service qui améliore l'existant et prolonge l'offre jusqu'au domicile, ainsi qu’une demande générale de flexibilité dans les couronnes d’agglomération. Ces services permettraient de capter une nouvelle clientèle, principalement des automobilistes, et donc de gagner des parts de marché. Les services proposés n'auraient donc pas uniquement pour effet d'améliorer la satisfaction des utilisateurs actuels (déjà conquis ou captifs ou dont les besoins sont en adéquation avec l'offre actuelle) mais représentent également une réelle opportunité pour CarPostal. Par ailleurs, on observe que les abonnements aux services de mobilité sont très intéressants, tant comme incitation à utiliser les services que comme source de revenus.
Download (.pdf)
Ce rapport présente les résultats d’une enquête quantitative menée dans le cadre d’un projet de recherche dans le domaine de la mobilité combinée, baptisé « Optima » et confié au Centre de Transport (TraCE) de l’Ecole Polytechnique... more
Ce rapport présente les résultats d’une enquête quantitative menée dans le cadre d’un projet de recherche dans le domaine de la mobilité combinée, baptisé « Optima » et confié au Centre de Transport (TraCE) de l’Ecole Polytechnique Fédérale de Lausanne. Ce projet de recherche vise à décrire et comprendre les pratiques de mobilités des personnes vivant hors des centres d'agglomérations afin de permettre de concevoir des solutions innovantes en matière de transport public. Cette recherche a consisté en une vaste enquête quantitative dont le but était d’évaluer les facteurs impliqués dans le choix modal des habitants. Cette étude, menée entre 2009 et 2010, ciblait une population d’habitants de communes périurbaines et rurales (20 000 personnes), parmi lesquels près de 2000 personnes ont répondu volontairement au questionnaire distribué (10% de la population-cible). Un aspect novateur de ce questionnaire est son développement autour de trois thématiques principales : les pratiques et habitudes de mobilité présentes et passées, les opinions ainsi que les attitudes face à la mobilité et enfin les données sociodémographiques des répondants et de leur ménage. A travers l’analyse combinée de ces informations, il s’agit de décortiquer les choix de modes de transport et de mieux comprendre les facteurs qui influencent ces choix individuels et, par la suite, faire apparaître les grandes tendances des pratiques et des perceptions des habitants des espaces d’intérêt de CarPostal. Car en effet, la mobilité des personnes ne se résume pas uniquement au choix des modes de déplacements, mais concerne plus largement les choix des modes de vie.
Download (.pdf)