ABSTRACT. Increased interactions between pedestrians and vehicles in current, crowded urban scenarios gives rise to a negative side-effect: a growth in traffic accidents, leaving pedestrians as the most injuried. Here, we combine public data sources, street level imagery and a deep learning to approach pedestrian and vehicle safety with an automated and simple data-processing scheme. The steps involved in this pipeline include the adaptation and training of a Residual Convolutional Neural Network to determine a hazard index for each given urban scene, as well as an interpretability analysis based on image segmentation and class activation mapping on those same images. The outcome of this approach is a fine-grained map of hazard levels across a city, a detailed analysis of the most influential urban objects in traffic safety and an heuristic to propose interventions to improve urban safety.

ABSTRACT. Cities exhibit different organizational patterns as a consequence of historical, political or economical circumstances, and constitute a paradigmatic example of complex system [1, 2]. In this context, network theory stands out as a fundamental tool facilitating the quantitative modelling of the main urban features and the analysis of the resulting dynamical processes, such as mobility and city growth. Our work focuses on street networks, which edges represent city roads, while the nodes portray the points where such roads cross. In the literature, the topic has been addressed following different points of views according the purpose and the system characteristics. For instance, the dynamics related to inter-urban roads (also known as arterial roads or high capacity roads), characterized by long segments and limited inter-connection, has been approached employing fluids models [3], or the fundamental diagram of traffic flow [4, 5]. On the contrary, phenomenology of intraurban streets is ruled by the underlying network structure and has been traditionally treated by means of graph models [6, 7, 8]. These two types of street networks, which are usually studied independently, are increasingly entangled as cities sprawl over suburban areas. So far, only few works dealt with street networks from an intertwined perspective, and the effects induced by such an interaction have been mostly overlooked. The analysis of the spatial interplay of intra and inter roads networks is the main motivation of the current abstract. We focus on monocentric cities and consider the situation in which arterial roads and urban local ones operate on separate geographic spaces. Specifically, local roads are located at the city center, and arterial ones at the urban periphery. Along this line, we develop the Grid-Tree model, or GT-model, where the former road frame is schematized as a square grid, while the latter one as a set of regular trees. The model reproduces previous results in terms of betweenness distribution [9] and, at the same time, offers a considerable advantage in terms of analytical tractability. GT-model allows to unveil several interesting properties of road networks with respect to the congestion phenomena. In particular, it evidences that cities may experience a set of multiple abrupt phase transitions in the spatial localization of congested areas. These transitions define a set of congestion regimes that correspond to the emergence of congestion in the city center, its periphery or in urban arterial roads, and regards the way in which different road classes are entangled to form a unique transportation system. The detection of congestion abrupt transition is performed both numerically and analytically and constitutes the main finding of our work [10]. The theoretical development we present is validated by looking into real road networks. Empirical analysis is carried out over almost a hundred of cities world- wide, and relies on automatic and unsupervised methods. Results show that the multiple abrupt transitions exist in real cities, confirming the prediction performed with the GT-model. The work provides another step forward the understanding of how urban growth and changes in mobility dynamics may affect urban road infrastructure.

ABSTRACT. Transport systems were organized to serve a specific set of citizens' needs, standing in time. However, those needs are rapidly evolving following technological, organizational and societal changes. Thus, a new and more dynamic way to manage transport systems is needed. Additionally, as more and more cities tend to be overcrowded, day-to-day problems like parking or congestion getting more complex and difficult to solve. Most of the studies estimate that around 70% of the world population will live in cities until 2050. As a consequence, static schedules of public transport (PuT) are not capable enough to accurately serve that dynamically changing demand. Nevertheless, a huge amount of passengers nowadays use smart-phone and feel comfortable to trust and adopt transport solutions, which based on mobile applications. As a result, of that unstable and dynamically changing ecosystem, transport systems should be oriented to be more flexible and demand-driven. This is why in the future PuT services need to consider Demand Transport Services (DRT) and ride-sharing (RS) as key solutions. Therefore, DRT and (RS) consist more as an idea and a theoretical method rather than a well-defined framework. Although most of the sub-problems involved in the final solution of planning a DRT system are well-studied, there is not a clear framework so far that leverage all of those ideas and provide clear guidelines to transport planners. In addition, transport planners use OD matrices which can represent mobility patterns in an acceptable level but they have some limitation when we search for decision tool on ride-sharing system planning. It’s more accurate to state that along with OD matrices that will reveal some patterns between days or maybe hours scale there is a need for an extra tool that allows visibility and knowledge extraction on trajectory terms. The current work aims to collect and synthesize some of the basic combinatorial optimization and statistical intelligence methods to give a more precise and elegant form on the planning of DRT and RS systems as well as the operational characteristics need to satisfy. The final purpose of authors is to compress with as less as possible parameters, that total system design and service features. In addition, the study propose to use of floating car data (FCD) to extract the origin and destination points of trips and use them as a proxy to analyze the suitability of DRT and RS systems.

ABSTRACT. During the last few decades, the complex systems theory has successfully disclosed the basic features of urban and regional dynamics. The time has come to make the next step, and apply our knowledge to controlling particular complex urban phenomena. The latter requires essentially deeper understanding of factors that govern these phenomena, verifiable behavioral models of human participants in these phenomena and representative data at high spatio-temporal resolution that adequately represent the heterogeneity of the urban space. Until very recently, the lack of the knowledge of human behavior and disaggregate data on urban physical space restrained urban complex system theory from becoming operational. We claim that the knowledge barriers are rapidly dissolving and apply complex system theory for practically-oriented modeling of urban parking dynamics. Parking is a perpetual problem of every big city. The imbalance between demand and supply in the central business and residential areas results in up to 30% additional urban traffic attributed to parking search. For the travelers, parking search time and price are among the major factors that define the choice of transportation mode and car ownership. For urban planners and managers understanding the dynamics of parking occupation pattern is critical for directing the inevitable transition from the nowadays city of automobile to the future Mobility as a Service. Parking is only loosely connected to the other components of the urban traffic and, thus, can be investigated separately from it. The major components of the system – parking demand and supply and drivers’ arrivals, departures and search behavior can be adequately estimated based on the standard high-resolution urban data that are available in the majority of the large Western cities. We thus consider the dynamics of the urban parking pattern is a unique example of a complex urban traffic subsystem that can be explicitly and formally represented at resolution of a single driver and single parking spot, and studied in theoretical and applied fashions. In the paper we investigate several analytical and simulation models of urban parking dynamics, all accounting for contemporary knowledge of drivers’ parking behavior. These dynamic are an outcome of the interplay between parking demand, supply and prices. Parking demand is created by residents and visitors, supply and prices are controlled by the city that aims at establishing a “proper” urban parking policy. We propose a set of models that fully captures urban parking dynamics and adequately forecasts the consequences of parking policy decisions in real cities. Our models demonstrate that parking dynamics exhibit all major attributes of a complex system – non-linearity, emergence and path dependence but nonetheless can be managed by means of the model-based pricing and soft constraints policy. Practically, we propose algorithms and software for estimating parking search time and establishing parking prices that guarantee a predetermined level of parking occupation. These applied solutions are built based on the standard high-resolution urban GIS data. We consider a variety of parking policies that improve the state of urban transportation and discuss the ability of local governments to implement these policies.

ABSTRACT. Please, see the extended abstract in attachment

ABSTRACT. The community structures are one of the most significant features of real-world networks. They provide a means to understand the complex interactions or relations (represented by network edges) between entities (represented by nodes) and offer an interpretable summary of a network. In parallel, the studies on migration flow have long been prominent in the fields of economics, sociology, geography, and policy decisions. This makes efficient modeling of migration flows, characterized by directionality, crucial to multiple disciples of research. This paper focuses on two relevant real-world migration flow complex networks- international asylum seekers migration flow and the inter-metro migration flow in the USA. Considering their directed and weighted nature, we explore different community detection methods and compare the performance of these methods on various quantitative metrics. We currently working towards identifying the issues in and improving the community detection methods for migration flows using directed network analysis.

ABSTRACT. The movements of individuals within and among cities influence key aspects of our society, such as the objective and subjective well-being, the diffusion of innovations, the spreading of epidemics, and the quality of the environment. For this reason, there is increasing interest around the challenging problem of flow generation, which consists in generating the flows between a set of geographic locations, given the characteristics of the locations and without any information about the real flows. Existing solutions to flow generation are mainly based on mechanistic approaches, such as the gravity model and the radiation model, which suffer from underfitting and overdispersion, neglect important variables such as land use and the transportation network, and cannot describe non-linear relationships between these variables. In this paper, we propose the Multi-Feature Deep Gravity (MFDG) model as an effective solution to flow generation. On the one hand, the MFDG model exploits a large number of variables (e.g., characteristics of land use and the road network; transport, food, and health facilities) extracted from voluntary geographic information data (OpenStreetMap). On the other hand, our model exploits deep neural networks to describe complex non-linear relationships between those variables. Our experiments, conducted on commuting flows in England, show that the MFDG model achieves a significant increase in the performance (up to 250% for highly populated areas) than mechanistic models that do not use deep neural networks, or that do not exploit geographic voluntary data. Our work presents a precise definition of the flow generation problem, which is a novel task for the deep learning community working with spatio-temporal data and proposes a deep neural network model that significantly outperforms current state-of-the-art statistical models.

ABSTRACT. Mobility patterns is a valuable information as it contributes to the efficient planning and operating of the transport networks. Demand in an area is usually estimated by sampling trips of people between each sub-area of a district and scaling up to the whole population. Nowadays, with the new technological, operational and societal changes, Origin-Destination (OD) matrices can be easily measured at every moment, so there is a need to define a methodology to fuse the set of OD matrices and to generate the representative ones for a concrete city. To generate there is a need to record trips, which are aggregated per zone in the OD matrix, which include information on the number of trips from every origin i to every destination j in every element OD(i,j) of theirs. The OD generation procedure is the second component in the traditional four-stage transportation forecasting model [1]. Through recording the start and ending points of each trip, a large volume of data is gathered which is then assigned to OD matrices for specific time intervals of each day [2] and specific days of the week. A logical and widely used choice for the merging of the smaller interval ODs towards producing a final OD is the statistical mean value of all the smaller interval ODs, averaging each OD (i,j) element. However, the mean value is a location estimate that assumes normally distributed values which is not always the case in practice. In this work, we propose the use of a new metric based on a novel local estimate for skewed data, the ε-med [3]. The two metrics are compared based on the conditional entropies of the final matrices with respect to the actual smaller interval ODs and the ε-med based one seems to outperform the classic mean value averaged one.

ABSTRACT. The modern information and communication technology in the 21st century provides vast and detailed data on human behavior. Until now, the human location data collection was so limited such as population surveys and spread of banknotes. But with the spread of mobile phones in recent years, it has become possible to follow the trajectory of individuals in detail. By analyzing such detailed observational data, many statistical properties have been reported. Recent studies of human mobility can be roughly categorized into two groups: One is focusing on statistical properties of individual behavioral patterns (microscopic scales), and the other is movement between the origin and the destination (macroscopic scales). However, studies of collective human flow (vector field) within scale of cities, which we call here mesoscopic scales, have been rarely addressed. In order to discover statistical properties of mesoscopic human flow patterns, we analyzed human flow as an analogy of river analysis, where the results were accepted for publication in Scientific Reports.

Our data is provided by Agoop, a Japanese company. The number of users is 260,000 people over a year, and GPS data consists of user ID, date, time, latitude, longitude, and velocity vector. It is collected at intervals of about 30 minutes, and the user ID is changed daily to protect privacy.

To characterize mesoscopic human flow patterns, we first divide the map into grids of 500 x 500m and calculate the average velocity vector of the moving people in each square in 30 minutes intervals. Next, we discretize the average velocity vector into four directions (east, west, north, south). For characterizing the flow patterns quantitatively, we introduce the concept of the basin that has been useful in the river flow pattern study. A square connects to an adjacent square according to the direction of the discretized velocity vector and considers the two connected squares to belong to the same basin. The basin is defined by adopting the same rules to all squares. Applying the method of basins to nine major cities in Japan (Tokyo, Osaka, Nagoya, Fukuoka, Sapporo, Sendai, Hiroshima, Okayama, Kumamoto), the size of the basin in the daytime is small, and its size distributions are similar to the simulation results when four directions are randomly selected. On the other hand, in the morning, a strong flow occurs toward the city center, and its size distribution can be approximated by a power law with an exponent of about -2.4. The distribution of number of moving people in basins in the morning follows a power law with the exponent close to -1.2. The difference between 2 exponents means that the basin area S and the population of moving people in the basin p are not proportional and have a non-linear relationship. Furthermore we define the diameter L of the basin as the maximum distance between two points in the basin and observe that p is proportional to L^3. This result contradicts the intuition that it is proportional to L^2 because of the two-dimensional map. It is speculated that this three-dimensional feature is related to the multi-layered structure in the city center due to skyscrapers and use of underground space. Also, a size S is also proportional to the diameter L to the power of 1.5. That is, the main transport link is characterized by a fractal structure of dimension 1.5. In this presentation, in addition to these scaling laws, we report that the perimeter P of the basin is proportional to the L^1.2. This scaling law indicates that the basin becomes more complicated as the expansion of urban areas.

Our results are strongly related to the fundamental parts of urban structure, such as the three-dimensional structure of cities and the expansion of urban areas. Currently, scientific studies on human mobility are attracting attention all over the world due to the spread of COVID-19. A new perspective on collective human flow patterns around big cities is applicable to all cities in the world and reveals the universal characteristics of human flow patterns.

ABSTRACT. Computational models that are able to integrate data from multiple sources can help to characterize the complex geographical and socioeconomic organization of population in urban areas with great resolution and applicability in practical contexts. This work presents a methodology for the construction of a complex system that represents different levels of the human's contact in an urban area. At a lower level, spatial information, obtained by geographical data combined to census data, is used to identify the representativeness of the population of each spatial region in the system (the demographic layer); a second layer, calculated from Call Detail Records (CDR) data, represents the social interaction between the individuals in the network and indicates how each region of the urban space interacts to each other (the social layer); a third layer, calculated from a combination between census and CDR data, represents the people's mobility in the urban space, indicating the Origin and Destinations of travels in the urban space (the mobility layer).

ABSTRACT. The advances of mobile technologies have led to massive amounts of data with respect to tracked routes of moving objects. The detection of anomalies in trajectory data is an evolving research domain, which has applications in traffic management and in public safety, but also in climate research and in animal habit analysis. It is a challenging task, due to the existence of non-linear spatiotemporal dependencies. In this work we propose the employment of an ensemble model, a combination of deep learning techniques with a traditional outlier detection methodology for detecting anomalous trajectories. We apply two variants of sequential denoising autoencoder architectures for unsupervised anomaly detection in vehicle trajectories, an LSTM-based autoencoder and a Sequence to Sequence LSTM autoencoder. A weighted distance-based loss function is optimized during the training phase, taking into account the impact of trajectory length on the detection outcome. We propose an ensemble architecture for the detection phase, which combines each of the autoencoders with the Local Outlier Factor algorithm, in order to detect anomalies. Our models are evaluated on different variants of synthetic anomalies. The results indicate a clear advantage of our approach with respect to performance.

ABSTRACT. While great emphasis has been placed on the role of social interactions as driver of innovation growth, very few empirical studies have explicitly investigated the impact of social network structures on the innovation performance of cities. Past research has mostly explored scaling laws of socio-economic outputs of cities as determined by, for example, the single predictor of population. Here, by drawing on a publicly available dataset of the startup ecosystem, we build the first Workforce Mobility Network among US metropolitan areas, and found that node centrality computed on this network accounts for most of the variability observed in cities' innovation performance. We attempted to generalize this paradigm to social media data by applying advanced NLP tools to online conversations with the aim of inferring the network of knowledge exchange between US states. We found that economic growth of a region is predicted much more accurately by the small fraction of social ties characterized by knowledge and trust than considering all ties indiscriminately. Our finding provides a first example of how open data and NLP could open up a novel way to model networks of exchange of non-material resources (such as knowledge) and to produce more nuanced, interpretable, and predictive network models of societal growth.

ABSTRACT. The Erasmus Program (EuRopean community Action Scheme for the Mobility of University Students), the most important student exchange program in the world, financed by the European Union and started in 1987, is characterized by a strong gender bias. Female students participate to the program more than male students. This work quantifies the gender bias in the Erasmus program between 2008 and 2013, using novel data at the university level. It describes the structure of the program in great detail, carrying out the analysis across fields of study, and identifies key universities as senders and receivers. In addition, it tests the difference in the degree distribution of the Erasmus network along time and between genders, giving evidence of a greater density in the female Erasmus network with respect to the one of the male Erasmus network.

ABSTRACT. Cities are microcosms representing a diversity of human experience. The complexity of urban systems arises from this diversity, where the services that cities offer to their inhabitants have to be tailored for their unique requirements. This paper studies the complexity of urban environments in terms of the assimilation of its communities. We examine the urban assimilation complexity with respect to the foreignness between communities and formalize the level of complexity using information-theoretic measures. Our findings contribute to a sociological perspective of the relationship between urban complex systems and the diversity of communities that make up urban systems.

ABSTRACT. We consider city as a system of networks: it is not a single system, but the coexistence of strongly interconnected systems that range from infrastructural to economic, from social to institutional supply chains, from the cultural fabric to urban planning. These dimensions are already a complex system, but their co-presence has generated one of the most complex systems in the world. Hannerz, in 1992, argued that all these levels contribute in defining the system of a city as "networks of networks" of relationships, groups, services, opportunities. Urban participation networks are part of this combination of flows and people, both at a micro, a meso, and a macro level: they are characterized by a complex social morphology since they are embedded in larger structures which can include several levels of action or layers of the urban social system. The tendency of citizens to form bonds of interpersonal collaboration depends not only on spatial co-habitation, but also on other types of relational micro dimensions to which they are affiliated in, and on the incorporation of these relational structures in a meso (groups) and macro (districts, neighborhoods) urban context. This kind of considerations have consequences not only in the ways in which people experience their environments, but also in individuals’ behaviours in terms of social interactions and groups, of sense of collective membership (community engagement), and of civic participation (i.e. informal and formal movements). The objective of this research is to understand the interpretative frameworks in investigating urban participation networks acting in a complex setting, both from an individual and a collective perspective. This abstract presents the research design – the theoretical discussion – behind an empirical study currently lead by researchers of the University of Pisa about the so-called Quartieri Uniti Eco Solidali (QUES) in the Tuscan city of Leghorn. QUES are neighborhood groups realizing a participatory path aimed at building an eco-friendly neighborhood model, based on a network of condominiums adopting practices of solidarity and environmental sustainability, and experiencing bottom-up forms of democracy. To understand participatory dynamics, we propose the use of a mixed-methods approach (Bellotti, 2015). In our study, we suggest a) to develop a multilevel analysis to investigate and measure the circular process occurring among relationships and civil participation, and b) the adoption of the typical qualitative analysis tools to deepen the construction of a collective identity and belonging (in-group and out-group mechanisms), which favour the development of community building processes and community empowerment. We discuss the analytical models that the theoretical-methodological perspective of networks gives in the field of sociological and urban studies; we complete our results referring to the theoretical and methodological approach of Symbolic Interactionism and Grounded Theory (Charmaz, 2006): if every action is a relationship, and is embedded in a collective dimension (the community, or the district, which are interactional settings), a non-standard analysis focuses on the “meanings” of networking and social participation. We highlight the conjunction of multiple aspects. The first one, more clearly, draws attention to the local community-participation relationship, understood as the embedding of participatory processes in data contexts (urban and suburban). They represent a key space where they assume a strategic function, both in terms of identification in symbolic and ideological-value aspects of which that possible social space is promoter, and in terms of the organisation and structuring of these forms of civic engagement (Giuffre, 2013; Lawlor & Neal, 2016; Neal, 2018). The second belongs to that debate where prevails idea that participation is an outcome of relational systems, whose plots are built and consolidated within the city. The interconnection of such patterns makes the urban system a system of networks of great complexity, so much so that it requires advanced methodological approaches analysed. Moreover, meanings people gives to belonging to the same neighborhoods (intended as informal groups, and movements) reinforce their sense of individual and collective identity, and the social roles they play in their everyday-life interactions (Mead, 1934; Goffman, 1959): they define their situation in a similar way (Berger & Luckmann, 1966), so they share collective symbols and values, and act in a similar way aligning their behaviours as a joint action (Blumer, 1969). As a result, neighborhoods are intended as a framework (Goffman 1974), in which participants (group members) share a common vocabulary of motives, and feel more engaged in their community because of their experience of group-values, which express membership processes.

From a methodological point of view, we choose to combine Social Network Analysis (SNA) measures with Grounded Theory methods. Science of networks has developed considerably in recent decades towards a better understanding of how urban spaces encourage the emergence, proliferation and embedding of forms of participation by citizens (Crossley & Diani, 2018). These forms of participation are the most evident manifestation of the weakening of traditional representative structures, i.g. political parties and associations, and of the development of new control mechanisms projected onto the physical and social spaces that citizens are close to (Della Porta & Andretta, 2001). Even though their less organised structures, these participative contexts develop from individuals need to find forms of personal commitment addressed to problems resolution linked to territories daily life (at a country level, neighbourhood, or often just street). Social Network Analysis allows to observe some phenomena recurring in urban social networks and to reconstruct structural configuration of these networks. Focusing explicitly on models of personal relationships, SNA theory and methods are intrinsically extra-individual and particularly appropriate to characterise social contexts. The tools offered by this approach – which make possible the empirical analysis of the link between social network attributes (dimension, composition, structure, flows) and territory - underline the importance of a procedural and unitary vision of social integration mechanisms, which cannot be understood as disconnected from contexts. Recent develops within this analytical perspective are Multilevel Exponential Random Graph Model (MERGMs), belonging to ERGM models family: analytical tools of great complexity, and potentially able to analyse multi-level networks in urban participation groups (Wang et al., 2013). The multilevel perspective is based on a sociological approach built on the idea that social contexts are not simply made up of "alters" understood as social actors, but rather of differentiated social foci around which shared activities are organised. Such social foci tend to induce dependency relationships among participants because: "individuals whose activities are organised around the same objective (focus) will tend to create interpersonal links, to form a group" (Feld 1981, p. 1015) and thus to condition each other (Crossley & Diani, 2019; Lomi & Stadtfeld, 2014).

Integrating this meso- and macro- approach with a study of the individual (micro) level, we are able to investigate nodes’ behaviours, or better the sense they give in participating in these kind of movements, such as QUES are: at the end, we would like to understand the sense people give to civicness as a joint action. Through participatory research, interviews and focus groups, and focusing on individuals social placement and role taking, we are able to reach a more complete analysis of processes of identity belonging: the asset on which these identity construction processes are based, and their role in implementing urban participation groups, and consequently their political/democratic success. If, according to Symbolic Interactionism perspective, every action is firstly an interaction and is embedded in a collective dimension, a deeper understanding of individuals’ meanings of network dynamics of collaboration can be helpful in discovering, analysing and developing the most useful directions for the empowerment of the community. And, by this, to improve bottom-up policies and community work, starting from actors’ perceptions about social problems, and what they suggest to do to face urban levels of criticality.

References

•Bellotti E. (2015). Qualitative Networks. Mixed methods in sociological research, London-New York: Routledge. •Berger P.L. & Luckmann T. (1966). The Social Construction of Reality. New York: Doubleday and Co. •Blumer H. (1969). Symbolic Interactionism. Perspective and Methods, Englewood Cliffs-New Jersey: Prentice Hall. •Bodin Ö. & Nohrstedt D. (2016), Formation and performance of collaborative disaster management networks: Evidence from a Swedish wildfire response, Elsevier, Global Environmental Change, 41, pp. 183-194. •Borlini B. (2010), Il quartiere nella città contemporanea. Come e perché occuparsene, Quaderni di Sociologia, LIV (52), pp. 13-30. •Brailly J., Favre G., Chatellet J. & Lazega E. (2016), Embeddedness as a multilevel problem: A case study in economic sociology, Social Networks, 44, pp. 319-333. •Brennecke J. & Rank O.N. (2016), The interplay between formal project memberships and informal advice seeking in knowledge-intensive firms: A multilevel network approach, Social Network, 44, pp. 307-318. •Charmaz K. (2006). Constructing Grounded Theory: A Practical Guide Through Qualitative Research. London: Sage Publications. •Coleman J. (1988), Social capital and the creation of human capital, Journal of Sociology, 94, S95-S120. •Collins C., Neal J. W. & Neal Z. P. (2014), Transforming Individual Civic Engagement into Community Collective Efficacy: The Role of Bonding Social Capital, American Journal of Community Psychology, 54, pp. 328 - 336. •Crossley N. & Diani M. (2018), Networks and fields in Snow, A. David, Soule, A. Sarah, Kriesi, Hanspeter, Mccammon, J. Holly (edited by), The Wiley Blackwell Companion to Social Movements, Oxford: Wiley Blackwell, 2018, pp. 151-166. •Della Porta D. e Andretta M. (2001), Movimenti sociali e rappresentanza: I comitati spontanei dei cittadini a Firenze, Rassegna Italiana di Sociologia, 42, pp. 41-76. •Feld S. L. (1981), The focused organization of social ties, American Journal of Sociology, 86(5), pp. 1015-1035. •Forrest R. & Kearns A. (2001), Social Cohesion, Social Capital and the Neighbourhood, Urban Studies, 38, 12, pp. 2125-2143. •Frank O. & Strauss D. (1986), Markov Graphs, Journal of the American Statistical Association, 81, 395, pp. 832-842. •Giuffre K. (2013), Communities and Networks. Using Network analysis to Rethink Urban and Community Studies, Polity Press, Cambridge. •Goffman E. (1959). Presentation of Self in Everyday Life. Garden City, NY: Doubleday Anchor. •Goffman E. (1974). Frame Analysis: An Essay on the Organization of Experience. Cambridge: Harvard University Press. •Gould R.V. (1991), Multiple networks and mobilization in the Paris Commune, 1871.Am. Sociol. Rev. 56 (6), pp. 716-729. •Hannerz U. (1992), Eplorare la città, Il Mulino, Bologna. •Henning C. & Lieberg M. (1996), Strong Ties or Weak Ties? Neighbourhood Networks in a New Perspective, Scandinavian Housing and Planning Research, 13, pp. 3-26. •Hollway J. & Koskinen J. (2016), Multilevel embeddedness: The case of the global fisheries governance complex, Social Network, 44, pp. 281-294. •Hunter A. (1979), The urban neighbourhood: its analytical and social context, Urban affairs, 14, pp. 276-88. •Lawlor J. A. & Neal Z. P. (2016), Networked community change: Understanding community systems change through the lens of social network analysis. American Journal of Community Psychology, 57(3-4), pp. 426-436. •Lomi A. & Stadtfeld C. (2014), Social Networks and Social Settings: Developing a Coevolutionary Vie, Springer Fachmedien Wiesbaden, 66, pp. 395-415. •Lusher D., Koskinen J.H. & Robins G.L., Eds. (2013), Exponential Random Graph Models for Social Networks: Theory, Methods and Applications, Cambridge, UK: Cambridge University Press. •Mead G.H. (1934). Mind, Self and Society from the Standpoint of a Social Behaviorist. Chicago, IL: University of Chicago Press. •Neal Z.P. (2018), Is the urban world small? The evidence for small world structure in urban networks, Networks and Spatial Economics. •Ohmer, M. L., & Beck, E. (2006), Citizen participation in neighbourhood organizations in poor communities and its relationship to neighborhood and organizational collective efficacy. Journal of Sociology and Social Welfare, 33(1), pp. 179–202. •Riccardo A. & Salvini A. (2020), La Social Network Analysis e le reti di partecipazione urbana in Partecipazione, Conflitti E Sicurezza. Mutamento e dinamiche evolutive dagli anni Cinquanta ad oggi (a cura di Elena Dundovich), Pisa University Press. •Robins G., Pattison P., Kalish Y. & Lusher D. (2007), An Introduction to Exponential Random Graph (p*) models for Social Networks, Social Networks, 29(2), pp. 173-191. •Salvini A. & Psaroudakis I. (2017), Dalle linee ai networks. Alcune implicazioni del pensiero di Blumer circa l’“allineamento dei corsi di azione”, Sociologia Italiana – AIS Journal of Sociology, N. X, pp. 191-200. •Salvini A., Riccardo A., Vasca F. & Psaroudakis I. (2020), Inter-Organizational Networks and Third Sector: Emerging Features from two Case-Studies in Southern Italy, in Ragozini G., Vitale M.P., Eds., Challenges in Social Network Research. Methods and Applications, Switzerland: Springer, pp. 209-233. •Van Den Berg P. & Timmermans H. (2015), A multi-level path analysis of social networks and social interactions in the neighbourhood, Region, 2, 1, pp. 55-66. •Wang P., Robins G., Pattison P. & Lazega E. (2013), Exponential random graph models for multilevel networks, Social Networks, Elsevier, 35 (1), pp.96 - 115. •Wang R., Tanjasiri S.P., Palmer P. & Valente T.W. (2016), Network Structure, Multiplexity, and Evolution as Influences on Community-based Participatory Interventions. Journal of Community Psychology, 44(6), pp. 781-798. •Wellman B. (1979), The community question: the intimate networks of East Yorkers, Am. J. Sociol. 84, pp. 1201-1231. •Zappa P. & Lomi A. (2016), Knowledge Sharing in Organizations: A Multilevel Network Analysis, in Lazega E. & Snijders T.A.B., Eds., Multilevel Network Analysis for the Social Sciences. Theory, Methods and Applications, Method Series 12. Switzerland: Springer, pp. 333-353.

ABSTRACT. See attachment.

ABSTRACT. In the aftermath of a natural disaster, a major challenge in coordinating relief operations is the lack of real-time information of resource-needs and resource-availabilities in the disaster-affected region. For disaster events that occur in urban regions, we propose to use Online Social Media as a source of such real-time information. We discuss some of the challenges in utilizing social media for coordinating post-disaster relief operations, and specifically address the challenge of mapping social media posts (microblogs) to resource classes as per UNOCHA guidelines.

ABSTRACT. Understanding how to create social tipping in urban systems, and the magnitude of change needed to make a difference in sustainability with implications on the regional components of the Earth system requires understanding of cross-scale phenomena, from citizen to neighbourhood and metropolitan area, from city to river basin, and from city to region. Complex systems science is a fundamental ingredient to model cross-scale phenomena. We will explore how to cross scales and showcase ongoing applications from economic experiments for consumers, producers and intermediate suppliers aiming to provide detailed understanding on how to trigger social tipping and change towards sustainability, and how to use experimental insights on models based on agents, stocks and flows, and networks at the neighbourhood and metropolitan scales. Then we will connect the metropolitan scale to growing scarcities under climate change and covid19 at river basin, region and global levels. Beyond systems understanding, our goal is to provide key input to the policy community, and a few examples of how (not) to communicate complexity to policy- and decision-makers will be shown and discussed.

ABSTRACT. Intertwined complex networks permeate every aspect of human activity. This complex- ity is especially unveiled in modern urban environments, which along with their merits bring several adversities. Imagining an emergency, such as the ongoing pandemic, arising from adversarial agents one can conceive the immediate threats posed to a highly dense and interconnected urban landscape. Consequently, it is vital to improve the design, robustness and resilience of the interdependent networks in question. Current emergency response strategies mainly rely on top-down management, e.g. policy makers trying to enforce behavioral policies on the population [2, 3]. However, bottom-up emergency responses may offer a valuable contribution to effective centralized emergency management, as demonstrated in the current pandemic, which saw groups of citizens spontaneously responding to provide relief to local authorities. Contributing to the aforementioned efforts, we model a specific aspect of emergency response, namely the formation of an accurate collective estimation of true danger (after some arbitrary event) in a specific geographical area, which is essential toward good emergency management. To that end, we employ an agent-based simulation where agents form a spatially distributed social network of shared danger estimates of their premises having limited samples. From the agents’ interactions we reconstruct the spatially distributed collective perception of danger and compare it to that of the true corresponding distribution in the model. Given a data-driven representation of the true danger, the model’s output can be a proxy of expected bottom-up responses to ongoing emergencies.

ABSTRACT. The COVID-19 pandemic caused an unprecedented global health crisis with high fatality rates that stressed the national health systems and the socio-economic structures of countries. National Governments have responded with non-pharmaceutical interventions (NPI) aimed at reducing the mobility of citizens to decrease the rate of contagion. This unprecedented scenario calls, indeed, for a better understanding of human mobility patterns during emergencies as well as in the immediate post-disaster relief. The study of mobility habits is a foundational instance for several issues ranging from traffic forecasting, up to virus spreading, and urban planning. The availability of rich datasets on the mobility of individuals, coupled with the urgency of the current situation, has fostered the collaboration between tech giants, such as Facebook and Google, institutions, and scholars. Along this path, we perform a massive analysis of aggregated and de-identified data provided by Facebook through its Disease Prevention movement maps \cite{maas2019facebook} to compare the effects of lockdown measures applied in France, Italy, and the UK in response to the COVID-19 outbreak. We focus on these countries due to the data availability at the time of the investigation. In fact, these are the only countries for which available data by Facebook cover also a period of time prior to the application of national lockdown. The overall dataset spans over 1 month of observations and accounts for the daily movements of over 13M people. We model countries as networks of mobility flow and we find that restrictions elicit geographical fragmentation through a transition toward local/short-range connections, thus causing a loss in the efficiency of mobility. Furthermore, to quantify the substantial effect of the lockdown we provide a model to simulate the effects of movement restrictions and find that the responses to the shock observed in real mobility networks can be fairly approximated through different network dismantling strategies. Indeed, the mobility restrictions caused a general reduction of the overall efficiency in the mobility network and a geographical fragmentation with a massive reduction of long-range connections. However, different countries experience changes depending upon their initial structure of inter-connections. The three countries exhibit differentiated mobility patterns that reflect the structural diversity in their underlying infrastructure: more centralized around their capital cities in the case of France and the UK, and more clustered in the case of Italy. Such infrastructural characteristics, together with different responses to national lockdown, contributed to the emergence of varied configurations in terms of residual mobility patterns. France shows one big cluster centered in Paris and many other smaller spots that disconnect as soon as the segregation process starts. Italy exhibits four interconnected clusters, centered approximately along the high-speed rail lines in Naples, Rome, Milan, and Turin, that remain interconnected over time thus showing a high persistence and resilience. Finally, the UK has one cluster centered in London, but most of England exhibits a higher persistence with respect to France and Italy, thus suggesting the presence of more capillary infrastructures. Our analysis confirms that the national resilience to massive stress differs and depends upon the inner connectivity structure. Understanding the resilience of mobility networks can contribute to enhance the preparedness to future systemic crises and to improve our predictive capabilities on the economic and social impact of mobility restriction policies.

ABSTRACT. The European Commission has identified that many member states are at risk of missing targets for the reuse and recycling of municipal waste. One approach to increasing household reuse and recycling that has shown promise is that of Pay-As-You-Throw (PAYT) or Save-As-You-Throw (SAYT) systems that impose costs on consumers for the production of undifferentiated waste or reward households for the correct separation of recyclable material. However, these systems still face challenges in achieving household participation, which has demonstrated to be suppressed to the low levels of trust in the waste management system. Here we propose the parachain blockchain as a promising solution to address these specific challenges. Typical blockchain solutions either suffer from issues with cost or scalability, in the case of public blockchains, or complexity, decentralisation, and transparency compromises in the case of private blockchains. Polkadot is a network of connected blockchains, called parachains, that provide “globally-coherent dynamic data structures”. Because parachains are linked through the Polkadot relay chain, they share a level of security without the requirement that each individual parachain achieves complete independent security. The development of a waste management parachain, as undertaken here, then offers the best of both worlds: the decentralisation and trust-less nature of a public blockchain with the customisation and scalability of an independent application specific blockchain. This under development parachain system will be incorporated into the BEE2Waste waste management system developed by Future Compta for a pilot project in the Oeste region of Portugal.

ABSTRACT. Innovative mobility services such as ride-sharing, bike sharing or micro-mobility services are rapidly flourishing in our cities. City planners need a way to analyse the planning and operation of such services in order to optimize their implementation from an economic and societal point of view. Four step transport modelling has been largely used to provide support to decision and policy makers, but the complexity of the new mobility services creates the necessity for lighter methodologies able to provide faster answers with less data requirements. Still, these will not replace conventional modelling approached, but be a first level analysis to provide initial insights and delimitate the analyses to be done at the next levels. The aim of this paper is to present analytical approaches for analysing the planning of ride-sharing, bike-sharing and micro-mobility services in order to optimize their operation in economic and social terms.

ABSTRACT. A system of hospitals in large cities could be considered as a large, diverse but interconnected system. Widely applied in hospitals, electronic health records (EHR) are crucially diverse from each other because of different health information systems, inside hospital rules, and personal behavior of physicians. The unstructured (textual) data of EHR is rarely used to assess the city-scale quality of health care. Within the study, we analyze the EHR data and in particular textual unstructured data as a reflection of the complex multi-agent system of city healthcare in Saint Petersburg. We describe available data for research and present a plan on how to develop a method for discovering a comprehensive record template using instruments which have been developed by our research group to process the natural language of medical texts.