Social Disorganization Theory (SDT) stated that in poor and unstable neighbourhoods, residents may have difficulty developing and maintaining social order, due to the weaknesses of their social networks and the infrequent exercise of informal control. As a consequence, in those areas criminal victimization tends to be high and persists over time. Latin American poor neighbourhoods are often characterised by high residential stability, dense informal networks, strong social cohesion, and yet they often have high levels of violent crime, which constitutes a challenge for SDT. Studies from new ecological approaches have asserted that even if informal networks are weak, neighbours can engage in actions to prevent crimes when the form of intervention is appropriately targeted and the activity is conducted in a partnership with agencies of public control, such as the police or local authorities. Thereby, the general distrust in police and local authorities, and the weak nexus between those institutions and local communities, which characterize most poor areas of Latin-American cities, represent relevant obstacles for the encouragement of neighbours' involvement in crime prevention initiatives. Despite the low rates of violent crimes in Chile, global figures tend to hide how complex the crime phenomenon is in the country, and particularly in Santiago city. In the capital and largest city of Chile, the distribution of High-Social-Impact crimes is highly unequal with a greater concentration of violent crimes in the most marginalized and poorest districts of the city. In this context is worth asking, to what extent do neighbourhood structural conditions, community-organizational mechanisms and new forms of public control influence the experiences of violent and property victimization in households of Santiago neighbourhoods? And, to what extent do such mechanisms mediate the relationship between structural conditions and the likelihood of being victim of a crime in Santiago neighbourhoods? To address these questions, the present study draws on an integral theoretical framework aimed at providing a holistic multilevel approach to explaining victimization risk across Santiago neighbourhoods. Data for this study are drawn from a community-survey of 5,860 persons (from 15 to 90 years old) who lived in 242 selected neighbourhoods of the Santiago city. The survey was conducted in 2010 by the Centre for Studies on Citizen Security (CESC), based at the University of Chile, in the context of their research project 'Crime and Urban Violence'. The hierarchical structure of the data (incorporating both individual and neighbourhood level measures) and the adaptation of internationally validated measurements, presents an excellent opportunity to evaluate complex hypothesis with advanced statistical tools. The research has shown that in neighbourhoods with a high concentration of poverty and low residential stability the probability of being a victim of violent crime is greater than in rich areas. However, when people manifest positive sentiments toward their neighbourhood, perceive collaboration and social cohesion among neighbours, and have positive perceptions with respect to police responses, this largely mediates the negative effects of structural conditions on household victimization by violent crimes, thereby eliminating these effects. These findings have important policy implications. They suggest that in disadvantaged communities it is imperative that police and local authorities not only try to reduce crime through traditional approaches, but also improve trust and engagement of the public aiming to build sustainable partnerships.

With the advent of smart sensor devices and Internet of Things (IoT) in the rapid urbanizing cities, data is being generated, collected and analyzed to solve urban problems in the areas of transportation, epidemiology, emergency management, economics, and sustainability etc. The work in this area basically involves analyzing one or more types of data to identify and characterize their impact on other urban phenomena like traffic speed and ride-sharing, spread of diseases, emergency evacuation, share market and electricity demand etc. In this work, we perform spatio-temporal analysis of various urban datasets collected from different urban application areas. We start with presenting a framework for predicting traffic demand around a location of interest and explain how it can be used to analyze other urban activities. We use a similar method to characterize and analyze spatio-temporal criminal activity in an urban city. At the end, we analyze the impact of nearby traffic volume on the electric vehicle charging demand at a charging station. / Master of Science / Because of the ubiquity of the Internet and smart devices, a tremendous amount of data has been collected from multiple sources like vehicles, purchasing details, online searches etc., which is being used to develop innovative applications. These applications aim to improve economic, social and personal lives of people through new start-of-the-art techniques like machine learning and data analytics. With this motivation in mind, we present three applications leveraging the data collected from urban cities to improve the life of people living in such cities. First, we start by using taxi trip data, collected around a given location, and use it to develop a model that can predict taxi demand for next half hour. This model can be used to schedule advertisements or dispatch taxis depending upon the demand. Second, using a similar mathematical approach, we propose a strategy to predict the number of crimes that can happen at a given location on the next day. This helps in maintaining law and order in the city. As our third and last application, we use the traffic and historical charging data to predict electric vehicle charging demand for the next day. Electricity generating power plants can use this model to prepare themselves for the higher demand emerged because of the increasing use of electric vehicles.