Atualmente empresas como Google e Facebook fazem parte da lista das maiores companhias do mundo. O investimento em publicidade e criação de páginas para a divulgação de anúncios e marcas, tem levado o Facebook a uma posição de destaque neste cenário. Neste contexto, o presente trabalho tem o objetivo de analisar e predizer o número de interações em notícias divulgadas em cinco páginas de fãs, que se constituem nas mais acessadas da Rede Social Facebook no Brasil. Como contribuição propõem-se determinar o fator de impacto de publicações, considerando a média de três características mencionadas, o número de curtidas, o número de comentários e o número de vezes que a notícia foi compartilhada. Serão avaliados resultados da aplicação de diferentes técnicas para a classificação, além da influência de características relacionadas a palavras e termos mais frequentes, verificando qual combinação produz melhores resultados no processo de gerar um modelo de aprendizado para prever o Fator de Impacto de notícias publicadas nas páginas de fãs da Rede Social Facebook. Apresenta-se também os motivos que podem exercer influência no fator de impacto através do processo de descoberta de conhecimentos em base de dados e também fazendo uso de técnicas de processamento de linguagem natural com o objetivo de atender a expectativa do trabalho. / Currently companies as Google and Facebook are on the top of the largest companies in the world and according to news released on the website tecmundo the main reason that led to this privileged position, in particular Facebook, appears to be the result of its investments in publicity focused on mobile devices through general advertisements in its own social network. In this context the present research aims to estimate the number of news interactions published on the five most accessed fans pages of Facebook Social Network in Brazil. Are considered examples of interactions in this study the number of likes, the number of comments and also the amount of times a message was shared. As also disclose attributes that influence interactions. As a contribution is proposed the impact factor of a publication, considering the average of three mentioned interactions, the number of likes, the number of comments and also the number of times the news was shared, in order to improve the results in predicting interactions of a fan page of Facebook Social Network. In addition to analyze the results of prediction algorithms applying different techniques of text pre- processing checking which combination produces best results in generating a learning process model to foresee the impact of news published on Facebooks fan pages and exhibit reasons that may influence the impact factor through the discovering process of database knowledge, from the feeling analysis as well as making use of processing of natural language techniques in order to fulfill work expectation.

No description available.

Crée en 1944, la Banque Internationale pour la Reconstruction et le développement entre en activité en 1947. Parmi les premières activités auxquelles elle se consacre au cours des années quarante, les missions et prêts de la Banque Mondiale en Colombie représentent un moment clef dans l’évolution de cette institution vers la forme d’organisation, l’articulation des procédures et la définition des objectifs qui la caractérisent encore aujourd’hui. Notre recherche porte sur ce processus, en se concentrant en particulier sur la mission organisée en 1949 et sur les conséquences de cette celle-ci sur l’évolution de la Banque, ainsi que sur la politique économique colombienne. Nous avons étudié l’ensemble des pratiques liées à l’évolution des politiques de la Banque mondiale dans les premières années d’activité, en soulignant en particulier le caractère émergent de cette organisation. Dans cette évolution, les missions représenteraient un moment d’interaction entre experts internationaux et experts locaux, fondamental pour comprendre l’évolution de l’action de la Banque mondiale. À partir de cette hypothèse, le défi principal de la recherche a été celui d’élaborer une méthode d’analyse permettant de saisir les stratégies et les pratiques mises en place au cours des missions de la Banque Mondiale et de les mettre en relation avec les résultats de ces missions. Pour saisir ces stratégies nous avons analysé les liens mises en place entre les acteurs, locaux et transnationaux, au cours des missions et en observer l’évolution dans le temps. La reconstruction de ces liens nous a permis de mieux comprendre l’émergence de certaines institutions à la base de la demande et de la gestion des prêts, ainsi que des critères d’évaluation et d’intervention des experts internationaux dans les pays en développement. Nous essayons donc d’étudier un phénomène transnational par une approche micro historique. / The International Bank for Reconstruction and Development was founded in 1944 and began its operations in 1947. Among the first activities the Bank carried out in the 1940s, the missions and loans in Colombia represented a key moment in the evolution of this institution towards the form of organization, the articulation of procedures and the definition of the objectives that characterize the Bank still today. Our research focuses on this process, focusing in particular on the General Survey mission organized in 1949 and its implications for the evolution of the Bank, as well as Colombian economic policy. Through this mission, we examined the practices related to the evolution of the World Bank's policies during the first years of activity, highlighting in particular the emergent nature of this organization. These missions, analyzed as a moment of interaction between international experts and local experts, are fundamental to understanding the evolution of the World Bank's work. Based on this assumption, one of the main challenge of the research was to develop a method of analysis to capture the strategies and practices developed during the World Bank missions and to relate them to the results of these missions. In order to understand these strategies, we analyzed the links established among the actors, local and international, and observed their evolution over time. Crafting and studying these links allowed us to better understand the emergence of certain institutions related to the demand and management of the loans, as well as the criteria the Bank established for the evaluation and intervention of international experts in developing countries. This implied the study of a transnational phenomenon by a micro-historical approach.

We have never stopped in the pursuit of science. Standing on the shoulders of the giants, we gradually make our path to build a systematic and testable body of knowledge to explain and predict the universe. Emerging from researchers’ interactions and self-organizing behaviors, scientific communities feature intensive collaborative practice. Indeed, the era of lone genius has long gone. Teams have now dominated the production and diffusion of scientific ideas. In order to understand how collaboration shapes and evolves organizations as well as individuals’ careers, this dissertation conducts analyses at both macroscopic and microscopic levels utilizing large-scale scholarly data. As self-organizing behaviors, collaborations boil down to the interactions among researchers. Understanding collaboration at individual level, as a result, is in fact a preliminary and crucial step to better understand the collective outcome at group and organization level. To start, I investigate the role of research collaboration in researchers’ careers by leveraging person-organization fit theory. Specifically, I propose prospective social ties based on faculty candidates’ future collaboration potential with future colleagues, which manifests diminishing returns on the placement quality. Moving forward, I address the question of how individual success can be better understood and accurately predicted utilizing their collaboration experience data. Findings reveal potential regularities in career trajectories for early-stage, mid-career, and senior researchers, highlighting the importance of various aspects of social capital. With large-scale scholarly data, I propose a data-driven analytics approach that leads to a deeper understanding of collaboration for both organizations and individuals. Managerial and policy implications are discussed for organizations to stimulate interdisciplinary research and for individuals to achieve better placement as well as short and long term scientific impact. Additionally, while analyzed in the context of academia, the proposed methods and implications can be generalized to knowledge-intensive industries, where collaboration are key factors to performance such as innovation and creativity.

L’identification des e-Leaders d’Opinion (e-LO) sur les réseaux sociaux internet est un enjeu important pour les managers marketing. Cette thèse s’intéresse à ces e-LO au travers de leur rôle, de leur statut et de leur identification. Chacun de ces trois piliers de la thèse fait l’objet d’une controverse dans la littérature. Les trois études menées dans cette thèse apportent des réponses complémentaires et croisées à ces controverses. Concernant le statut de l’e-LO, nous définissions l’e-LO dans les réseaux sociaux internet selon trois caractéristiques : l’expertise dans la catégorie de produit, la volonté de publier des contenus valorisables pour leurs contacts et la volonté de développer leur capital social. Concernant l’identification des e-LO, une échelle de mesure multidimensionnelle dotée d’une bonne validité externe est développée. Elle est confrontée à des mesures d’e-LO et de LO inspirées de la littérature, mais également à des mesures alternatives comme la métrique algorithmique Klout et des indicateurs sociométriques de profil Twitter. Enfin, quatre modèles et leurs variantes sont testés pour mieux comprendre l’origine et le mécanisme de l’e-influence de l’e-LO (rôle de l’e-LO). Nous montrons que l’e-LO a un impact sur l’e-influence dans la catégorie de produit. La confiance joue le rôle de médiateur entre l’e-LO et l’e-influence. Ces résultats plaident pour l’intérêt des e-LO pour le marketing viral. / E-Opinion Leaders (e-OL) identification on internet social networks is a major issue for marketing managers. This thesis deals with the role, status and identification of these e-OL. Each one of these three aspects are subject to controversial debates in the literature. Three studies are conducted within this thesis. They provide complementary and crossed answers to these debates. Regarding the e-OL status, we offer a definition for the e-OL concept in the specific context of internet social networks within three main characteristics: expertise in the product category, the will to publish valuable contents for their contacts, and the will to develop their social capital. Then, for the e-OL identification, we develop a new and multidimensional measurement scale for this concept. This scale shows good external validity. We confront it with others e-LO measures inspired by traditional OL measures, but also with alternative identification methods as the Twitter profile indicators and the Klout score algorithm. Finally, we test four models and their variants in order to better understand the origin and mechanism of the main e-LO role: the e-influence. We especially show that e-OL has an impact on e-influence. We also show that confidence mediates the link between e-LO and e-influence. These results speak to the relevance of e-OL in a viral marketing perspective.

During the transition to adulthood, youth often rely on the people in their life for support. However, for transition-aged foster youth, these supports may not be available or may look different than supports available to youth in the general population. Relationships with supportive adults have been found to help former foster youth transition to adulthood, but little is known about the adults youth have in their network. Foster youth who report increased levels of social capital have been shown to experience higher levels of success in young adulthood. However, as former foster youth transition to adulthood, a lack of in-depth understanding of supportive adults and social networks creates difficulties identifying—and addressing—potential gaps in their social network. This study aims to gain a better understanding of how social networks influence social support and access to and mobilization of social capital as youth leave the foster care system. A social network assessment based on two existing measures was created to attain a better understanding of the social networks of transition-aged foster youth. The new social network tool was piloted with a group of young adults prior to use in this study. This social network tool allowed for an in-depth understanding of social networks, social support, and social capital as three distinct constructs. The social network characteristics included: on whom the youth relies for support, how the relationship developed, and the closeness of the relationship. Social support included: questions on the type of support available to youth (resources, emotional, advice, or constructive criticism), as well as the social support domains (housing, education, employment, relationships, and transportation). Social capital was examined based on questions about network members’ occupation(s) and frequency of communication between the youth and each network member. Univariate, Bivariate, and Multivariate analyses were utilized to examine social network characteristics, foster care history, social support, and social capital. The mean network size of sample participants was 7.1 and the range was 0–36. A relationship between placement type and social network members was found; indicating that youth in congregate care were more likely to have formal (social service related) networks than youth not in congregate care. A relationship between having more informal network members and housing stability was found; indicating that different network members may help youth with different young adult outcomes. A relationship between both access to, and mobilization of, social capital was found based on the type of social network members (formal, informal, familial-biological, familial-foster). Based on the findings of the current study, research and practice implications are discussed. These include the utility of social network analysis with transition-aged foster youth, future lines of inquiry, and suggested practice/policy shifts.

This study develops a new theoretical approach and empirical measure of American regional subcultures using public opinion survey data and building on previous research (Chinni and Gimpel 2011; Elazar 1962, 1966; Hero 2000; Lieske 1993; Putnam, Leonardi and Nanetti 1994). Instead of approaching classification of regions based on formal geography, border states, population demography, ethnic groups and migration patterns, or historical traditions, this study uses a vernacular geography approach to study culture in the 50 American states. Vernacular geography is the sense of place revealed in ordinary people’s language. The study uses original nationwide survey data to measure perceptions of place based on states that are most similar to a respondent’s home state. The measure is based on unique survey questions where respondents have the freedom to choose any of the 50 states. The surveys are conducted by the Cooperative Congressional Election Study (CCES) from 2012 to 2016. These data allow development of a new measure of state similarity or regional subcultures based on vernacular geography. The state similarity network based on people’s feelings shows that state contiguity is not the driving factor in people’s perceptions of regions of the United States. It also shows that people’s perceptions of state similarity are a better predictor of policy diffusion than contiguity. Finally, this study shows that wealth is the most important factor in people’s perceptions of state similarity, but that population size, racial diversity, rural/urban population density, and ideology/partisanship are all predictors of people’s perceptions of state similarity at low levels. This study argues that perceptions of place matter. They are a core building block of political culture and are important for understanding policy diffusion. This study is about how citizens conceptualize their home state and network of most similar states, and whether state similarity networks, or social networks of states, influence government policy adoption and innovation.

Social media has changed the way that we create, use, and disseminate information and presents an unparalleled opportunity to gather large-scale data on the networks, behaviors, and opinions of individuals. This dissertation focuses on the role of social media and social networks in recruitment, examining the complex interactions between offline recruiting activities, online social media, and recruiting outcomes. Specifically, it explores how the information college football recruits reveal about themselves online is related to their decisions as well as how this information can diffuse and influence the decisions of others. Recruitment occurs in many contexts, and this research draws comparisons between college football and personnel recruiting. This work is one of the first large-scale studies of social media in college football recruiting, and uses a unique dataset that is both broad and deep, capturing information about 2,644 recruits, 682 schools, 764 coaches, and 2,397 current college football players and tracking offline and online behavior over six months. This dissertation comprises three case studies corresponding to the major decisions in the football recruiting cycle—the coach’s decision to make a scholarship offer, the athlete’s decision to commit, and the athlete’s decision to decommit. The first study investigates the relationship between a recruit’s social media use and his recruiting success. Informed by previous work on impression management in personnel recruitment, I construct logistic classifiers to identify self-promotion and ingratiation in 5.5 million tweets and use regression analysis to model the relationship between tweets and scholarship offers over time. The results indicate that tweet content predicts whether an athlete will receive a new offer in the next month. Furthermore, the level of Twitter activity is strongly related to recruiting success, suggesting that simply possessing a social media account may offer a significant advantage in terms of attracting coaches’ attention and earning scholarship offers. These findings underscore the critical role of social media in athletic recruitment and may benefit recruits by informing their branding and communication strategies. The second study examines whether a recruit’s social media activity presages his college preferences. I combine data on recruits’ college options, recruiting activities, Twitter connections, and Twitter content to construct a logistic classifier predicting which school a recruit will select out of those that have offered him a scholarship. My results highlight the value of social media data—especially the hashtags posted by the athlete and his online social network connections—for predicting his commitment decision. These findings may prove useful for college coaches seeking innovative methods to compete for elite talent, as well as assisting them in allocating recruiting resources. The third study focuses on athletic turnover, i.e., decommitments. I construct a logistic classifier to predict the occurrence of decommitments over time based on recruits’ college choices, recruiting activities, online social networks, and the decommitment behavior of their peers. The results further underscore the power of online social networks for predicting offline recruiting outcomes, giving coaches the tools to better identify vulnerable commitments.

Social networks encode important information about the relationships between individuals. The structure of social networks has important implications for how ideas, information, and even diseases spread within a population. Data on online social networks is becoming increasingly available, but fine-grained data from which physical proximity networks can be inferred is still a largely elusive goal. We address this problem by using nearly 20 million anonymized login records from University of Iowa Hospitals and Clinics to construct healthcare worker (HCW) contact networks. These networks serve as proxies for potentially disease-spreading contact patterns among HCWs. We show that these networks exhibit properties similar to social networks arising in other contexts (e.g., scientific collaboration, friendship, etc.) such as the "Six Degrees of Kevin Bacon" (i.e., small-world) phenomenon. In order to develop a theoretic framework for analyzing these HCW contact networks we consider a number of random graph models and show that models which only pay attention to local structure may not adequately model disease spread. We then consider the best known approximation algorithms for a number of optimization problems that model the problem of determining an optimal set of HCWs to vaccinate in order to minimize the spread of disease. Our results show that, in general, the quality of solutions produced by these approximations is highly dependent on the dynamics of disease spread. However, experiments show that simple policies, like vaccinating the most well-connected or most mobile individuals, perform much better than a random vaccination policy. And finally we consider the problem of finding a set of individuals to act as indicators for important healthcare related events on a social network for infectious disease experts. We model this problem as a generalization of the budgeted maximum coverage problem studied previously and show that in fact our problem is much more difficult to solve in general. But by exposing a property of this network, we provide analysis showing that a simple greedy approach for picking indicators provides a near-optimal (constant-factor) approximation.

This research offers an analysis of two concepts: subjective well-being and social capital. I examine each concept separately, looking at their underlying determinants and characteristics, and I also look at the links between these two notions, as there is theoretical support for their interdependent relationship. I have conducted data analysis of five samples, all part of the World Values Survey 1995-1996. The WVS 1995-1996 offers vast data for Spain and for four autonomous regions - Basque Country, Andalusia, Galicia and Valencia. For the national sample N= 1211. Regional sampling is as follows: Basque Country (N=2205), Andalusia (N=1803), Galicia (N=1200), and Valencia (N=501). The subjects of all the surveys are citizens of both sexes, aged 18 and older.