COST TU1305 Workshop in Mobile Tartu  |  Summary by Dr. Slaven Gasparovic

COST TU 1305 Workshop in Mobile Tartu 2106

29.6.16 Tartu, Estonia

Summary by Dr. Slaven Gasparovic

 

 

As  part of the International Scientific conference Mobile Tartu 2016, COST TU1305 workshop and PhD seminar, with a field experiment of smartphones and GPS tracking, took place on 29th of June 2016 in Järveveere holiday centre, near Tartu.

Prof. Rein Ahas (Tartu University, MC member and WG1 leader of COST TU1305) opened the conference and introduced leaders of the working groups of the Field Experiment as a part of PhD seminar.

Keynote speaker was Prof. dr. Stefan van der Spek from TU Delft. In  his lecture ''Sensing the City/Advances in Tracking Technologies'', he discussed the relation between movement and the cognitive system, the relation between actual movement and structure offered. One of the most important questions within his speaking was about possibility to reveal the logic of people's movement and translate this into design of smarter cities. For this purpose social media and VGI are becoming more and more important sources for understanding perceptions and experiences. The built environment provides a framework for our movement. This framework is on a longer scale temporal as well as resulting in both spatio-temporal data and temporal-spatial data. The future is using smartphones and user-generated data for (parametric) design and interventions.

 

Field Experiment of smartphones and GPS tracking, as a part of PhD seminar, was organized in three parts and in three groups.

Part 1 - 13:00 - 13:30. All participants chose to participate in one of the groups:

-       Group A ''Smartphone tracking & social network analyses with YouSense'' (leaders Anto Aasa, Anniki Puura),

-       Group B ''GPS tracking & urban planning'' (leader Stefan van der Spek)

-       Group C was ''CDR data exercise'' (leaders Erki Saluveer, Margus Tiru).

In this first part leaders of the groups explained to participants all the tasks, tactics and gave the data/devices/materials to participants to work with.

Part 2 - 19:10 – 20:00 All participants, i.e. fieldwork groups continued with their work from 19:10 until 20:00. Some participants worked with the GPS devices, some of the participants analyzed the data etc. 

Part 3 - Next day (30th of June) all groups presented their work within the separate session at the end of the day.

 

COST TU1305 seminar was organized in two sessions, chaired by Prof. Rein Ahas (from 14:20 until 16:00) and Prof dr. Joao de Abreu e Silva (from 16:20 until 18:00). In total 10 lectures were held with very good feedback from other participants (in average 2-3 questions after every lecture).

Thibaut Dubernet (ETH Zurich) presented a topic regarding new data for simulating the impact of social networks on travel. The lecture aims at filling a data gap that exists in data to better understand the way individuals plan activities together, and how much this kind of joint process might be relevant for transport planning. Based on the needs of a simulation-based experimentation, a two-fold data collection is designed (Call Detail Records and an interactive survey).

Maria del Mar Alonso-Almeida (Autonomous University of Madrid, MC Substitue in COST TU1305) performed a lecture regardingsocial media use in mobile devices and leisure tourism travel behaviour. The main aim was to understand the link between the social networks and apps' use in leisure travel on mobile devices between Portugal and Spain, to create a measure tool in order to collect information, to perform a survey and analyze a data and to provide information to main stakeholders. She concluded that social media are mainly used in mobile devices in a multiple way with relevant role of Instragram and with clear differences between Portugal and Spain (her work is based on STSM she has accomplished in COST TU1305)

Arnaud Adam(Université catholique de Louvain, Young Researcher in COST TU1305) presented a topic regarding revisiting the urban structure of Brussels by means of mobile phone data and community detection. The objective of his lecture was to detect the socio-spatial organization within the metropolitan area of Brussels by means of almost 20 million mobile phone calls collected by one operator. Using the origin and the destination of all calls, the Louvain method is applied to detect communities of places. He showed clear spatial pattern in and around Brussels, phone data added an interesting viewpoint to more classical urban models and methodological choices impact the delineation and the composition of the metropolitan area which questions their use for decision-makers.

Emmanouil Chaniotakis (Centre for Research and Technology Hellas-Hellenic Institute of Transport, MC Substitute in COST TU1305) discussed about understanding activity patterns from social media data (a fusion of Facebook and Twitter). He presented a research about activity patterns found from geo-referenced data coming from social media, by analysing and comparing data collected from two major social media websites (Twitter and Facebook) in the same location and period of time. Data obtained from both sources is analysed and fused based on its temporal and spatial distribution. Data from Facebook (check-in events) is used to define Points Of Interest (POIs), while data from Twitter is used to define the activities performed. His presentation emphasized that most of the check-ins performed (Facebook) and the tweets posted (Twitter) are related to leisure activities, and that there is high spatial correlation among the two different sampled data.

João de Abreu e Silva (University of Lisbon, MC member in COST TU1305) presented the relationship between travel behavior, ICT usage and social networks in university students in 3 European cities. He presented the results obtained from a web-survey applied to university students in 3 southern European cities, Lisbon, Granada and Zagreb. The survey was aimed at studying the relations between ICT and social media usage, social networks and social travel in this specific sociodemographic group. It aims to investigate a series of potential research questions related with the effect of ICT on social travel, its interaction with social networks, and, how the perception of its usefulness influences it usage, and to research the existence of differences between these three places and examine the role of cultural differences in explaining them. A description of the obtained results is presented as well as a series of statistical tests related with the research questions that oriented the survey design.

The topic of Josep Maria Salanova Grau (Centre for Research and Technology Hellas-Hellenic Institute of Transport) was about social media for transportation planning (estimating demand for taxi). He presented the study which investigates practical applications of using social media data for enhancing the transportation planning capabilities. A methodological framework is defined for the identification of the citizens’ mobility patterns using social media data. The data, composed by the geographic location of the check-in locations and the time intervals in which the data is extracted (in aggregated values) is obtained by Facebook through the check-in service. The results illustrate the actual correlation of demand for taxis to the check-in events, while it is noted that models should take into account the hysteresis of the demand production to the check-in events for departure from the location of the check-in based on the type of location. He discussed practical implications of the findings and the model derived.

Domokos Esztergar-Kiss (Budapest University of Technology and Economic, MC member in COST TU1305) performed a lecture regarding the extensions of the activity chain optimization method. He elaborated a method for the optimization of daily activity chains. Flexible demand points were introduced, because some activities are not necessarily fixed temporally and spatially, therefore they can be realized in different times or locations. The simulations were performed on arbitrarily chosen test networks using Matlab. He presented promising results for all transportation modes (car, public transport, combined). The application of the extended method produced shorter activity chains and a decrease of travel time for travelers. 

Stefano Pensa (The Polytechnic University of Turin, MC Substitute in COST TU1305) presented a topic regarding a map comparison for the evaluation of usability of Twitter data (traditional vs data-driven approaches). He described the use of quantitative data obtained from the social media Twitter to analyse urban patterns. A case study in the urban neighbourhood of Pampulha,  Belo Horizonte, Brazil, was described to show the outcomes of a comparative analysis with traditional data used in the urban and transport planning processes of the city. He showed similarities and differences to evaluate the accuracy of analysis performed by the use of freely and quickly available data instead of costly but official data.

Oded Gal (Technion - Israel Institute of Technology) discussed about shared-private transportation and social media: the role of social media as platform for ridesharing (developing an alternative model). He showed a study which attempts to examine how people's perception of ridesharing influence their willingness to use it as a model of urban transportation. The study is based on a questionnaire. The main question was how people perceive the trust factor in social media ridesharing platforms. He showed two models for the integration of ridesharing in social media: dynamic dominant and trust-dominant.

Adeline Decuyper (Université catholique de Louvain) presented the topic about the issues of spatial aggregation for community detection in communication networks. The objective of her research was to address specific methodology issues that arose from the nature of large datasets collected automatically. As case study, she analyzed the communities within the Brussels metropolitan area, using Twitter and mobile phone data. Regarding results she showed that pre-aggregating the network by geographical areas can have a high impact on the structure of the communities that are found. She discussed the biases that can be present in the datasets analyzed, and proposed methods to compensate for those biases.

 

At 18:00 we gather to dinner of some food specialities from Estonia. The 2nd part of the Field Experiment followed dinner and ended the Conference’s first day.

During the evening a social program was organised. All participants had a chance to try sauna, swimming in a pool or walking. Everybody went back with a special bus to Tartu for the 2nd day of the Conference.

Mobile Tartu 2016 Conference’s first day including COST TU1305 Session and Field Experiment was organized in a beautiful environment on the lake bord of Järveveere holiday centre. Prof. Rein Ahas and all his crew gave their best and hosted all participants excellent. Field Experiment and COST TU1305 Session was very usufel to get some new knowledge and insights about topics regarding ICT, social network and travel behaviour.