Purpose: The purpose of this study is going to visualize the public sentiment on expected and unexpected social events. Exploring the relationship between tweets forwarding and sentiment. Design/methodology/approach: This research related to sentiment analysis of social events applied a lexicon-based method. The social events come from Facebook data breach and Ireland vote on abortion event. The study conducted This study focused on how the public sentiment changes over time and the relationship between sentiment and tweet forwarding. Bing lexicon and NRC lexicon are adopted in the analysis. Result: The result of this study is the dominant sentiment trend is consistent with the trend of the number of tweets over time in the Facebook data breach and Ireland vote on abortion. Besides, the sentiment has affected people forward tweets in this research.

abstract: The purpose of this study was to examine the local residents' perceptions toward the social impacts of FIBA EuroBasket 2011, held September 7th to September 12th, 2011 in Vilnius, Lithuania. This study focuses on the social impacts of the event since most of the mega-event academia is dominated by economic impact studies and this research attempts to go beyond a narrow quantitative approach to examine the social impacts in relation to two important academic contributions: 1) the embracement-withdrawal continuum and 2) the social exchange theory. A mixed methods analysis was adopted as the best approach for this research. Both a quantitative survey and qualitative questionnaire were used which yielded a total usable sample of 128 residents of Vilnius. Exploratory Factor Analysis was conducted to examine the underlying themes of the quantitative data and coding and thematic development was used to make sense of the qualitative data. The results confirmed that the event was considered successful (97% of residents embraced the event) and they generated 6 factors, or impacts, that were deemed important for influencing resident perceptions: social costs, social benefits, sociocultural impacts, socioeconomic impacts, sociopolitical impacts, and socioenvironmental impacts. These results indicate that the use of mixed methods analyses is crucial to investigate the interdependence between each impact that affects residents' perceptions toward sport mega-events. / Dissertation/Thesis / M.S. Recreation and Tourism Studies 2011

We present a theoretical approach to the statistical analysis of the dependence of the gap time length between consecutive recurrent events, on a set of explanatory random variables and in the presence of right censoring. The dependence is expressed through regression-like and overdispersion parameters, estimated via estimating functions and equations. The mean and variance of the length of each gap time, conditioned on the observed history of prior events and other covariates, are known functions of parameters and covariates, and are part of the estimating functions. Under certain conditions on censoring, we construct normalized estimating functions that are asymptotically unbiased and contain only observed data. We then use modern mathematical techniques to prove the existence, consistency and asymptotic normality of a sequence of estimators of the parameters. Simulations support our theoretical results.

Crisis communication has been widely researched for an economic purpose, with the focus on preventing damage to a corporation’s reputation or the restoration of a politician’s image. Additionally, crisis communication also found its utility for natural hazards as well as for terrorist attacks. However, identified as being a research gap, this thesis focuses on studying the prepared crisis communication of mega-event and festival organizations in case of a serious incident. The purpose of this thesis is to contribute to the area of crisis communication in the way that it focuses on the process of preparing for a possible crisis with the creation of effective communication manuals, action plans, etc. rather than investigating the post-crisis communication. Through using the uncertainty reduction theory and the chaos theory, it is possible to investigate the internal crisis management processes of organizers that go beyond protecting its reputation, but focus more on saving human lives as uncertainty and chaos get reduced. A deductive approach was taken through conducting interviews and carrying out a thematic analysis of written documentary sources, such as manuals and action plans, as well as the transcripts. In general, one can say that the awareness about the importance of a crisis management and communication plan rises. Regular trainings and a constant collaboration with external partners, such as the blue light organizations, are seen as central to the emergency planning process. Moreover, to guarantee an effective crisis communication, often smaller task groups within the crisis management team are in charge of instructional crisis communication. Knowledge about different threat-scenarios and functions of external specialists reduces uncertainty and supports actions to reduce chaos. This thesis and its findings have a societal relevance, in the sense that it allows a wider and deeper understanding of crisis communication in an area of application that does not seem to be very common. Through insights into the crisis management process of the different event- and festival organizations, one can become active today and threats to human lives can be minimized or even avoided in the future.

La structure du milieu interstellaire est très hétérogène dans notre Galaxie. Les observations, ainsi que les lois d'échelle, suggèrent que sa structure pourrait être fractale. Les plus petites structures de gaz ou clumps auraient un rayon de quelques unités astronomiques et une température de quelques kelvins, les rendant très difficilement détectables. Ces nuages de gaz moléculaires froids et denses peuvent donc être des candidats peu lumineux à une matière noire baryonique.L'objectif de cette thèse est de trouver des indices observationnels de l'existence de ces sous-structures gazeuses. La méthodologie consiste à comparer les magnitudes apparentes d'étoiles à différents intervalles de temps et vers plusieurs directions dans la Galaxie; afin de détecter des variations photométriques temporelles compatibles avec un événement d'obscurcissement. En effet, si un nuage moléculaire se trouve en amont d'une étoile, celle-ci sera obscurcie ou occultée par les sous-structures composant le nuage moléculaire, pendant une certaine durée de temps (dépendant de la vitesse de chaque objet et du temps d'observation considéré). Dans un premier temps, nous avons effectué un traitement photométrique complet pour chercher de tels événements dans l'amas globulaire NGC 4833 en comparant des observations photométriques séparées de six mois. Cet amas fut observé par D. Puy et D. Pfenniger en janvier et juillet 2006. Puis, nous avons utilisé des observations de plusieurs amas globulaires observés par le télescope Hubble, dont la photométrie est connue et rendue publique.Finalement, nous avons développé des méthodes d'apprentissage automatique afin de détecter des événements d'obscurcissement dans le Sloan Digital Sky Survey, Stripe 82.Aucun événement d’obscurcissement d'étoiles n'a pu être détecté ni dans la direction des amas globulaires considérés ni dans la région de Stripe 82. Néanmoins, des simulations d'un nuage fractal ont été réalisées pour estimer la probabilité d’occurrence du phénomène d'obscurcissement d'étoiles. Elles montrent que l’événement est extrêmement rare, ce qui permet d'émettre une nouvelle stratégie d'observation pour augmenter les chances de détecter de tels événements. De plus, les méthodes d'apprentissage automatique utilisées pour cette étude sont appliquées à d'autres objets variables tel que les quasars, très nombreux dans Stripe 82, et peuvent être utilisées pour d'autres objets variables tels que les supernovae. / The scaling laws and the very appearance of the interstellar medium suggest that this medium might have a fractal structure. The smallest structure of gas called clumps might have a radius about few astronomical units and a very low temperature of few kelvins, making them very difficult to detect. That is why, cold, dense, molecular clouds could be ideal candidates for baryonic dark matter. Because of their low temperature, they do not radiate, and seem to be dark.The aim of this thesis is to highlight observational clues on the existence of clumps by comparing apparent magnitudes of stars at different times, and towards different directions in the Galaxy. If a molecular fractal cloud lies behind a star, it will be obscured or occulted within a certain time (depending on the velocity of each object and on observational time). Therefore the goal of this thesis is to detect magnitude-time variations of stars compatible with an obscuration event.First, we did a complete photometric data reduction of observations of a globular cluster, NGC 4833, carried out by D. Puy and D. Pfenniger in January and July 2006. We wanted to detect obscuration events of stars in NGC 4833, in a six-month period. Then, we used public data of several globular clusters observed by the Hubble Space telescope.Finally, we developed machine learning methods to detect obscuration events in the Sloan Digital Sky Survey, Stripe 82.No obscuration events were detected either in several globular clusters, or in the Stripe 82 survey. However, we performed simulations of a fractal cloud to estimate probabilities of obscuration events. They show that such events are very rare, allowing us to develop a new observational strategy for increasing the probability of detecting obscuration events. Moreover, machine learning methods used in this study are applied to other variable objects, such as quasars which are numerous inside Stripe 82 survey, and could be used to other variables objects such as supernovae for example.

L’étude vise à caractériser les indicateurs relatifs aux évènements hydroclimatiques extrêmes dans le bassin-versant de l'Ouémé à l'exutoire de Bonou. Pour atteindre cet objectif, les données climatologiques (pluies journalières, températures journalières, évapotranspiration potentielle journalières), les données hydrométriques (débits journaliers), les données de réanalyses (pression atmosphériques, humidité relative, vitesse du vent, Indices de Mousson West Africain), la température de surface océanique sur la côte de Cotonou et les données cartographiques relatives aux occupations du sol des années 1986, 2000 et 2015 sont utilisées. L’analyse de la variabilité hydroclimatique a révélé que le secteur d’étude a connu sur la période de 1951-2015, deux ruptures de stationnarité (1968 et 1987) qui ont divisé la série en trois sous-périodes. La variabilité pluviométrique révèle qu’il une régression de 18 % (au sud), de 16 % (au centre) et de 15,8 % (au nord) entre la première sous-période (1951-1968) et la deuxième sous-période (1969-1987), alors que la dernière sous-période (1988-2015) est marquée par une reprise pluviométrique de 15 % (au sud), de 16,3 % (au centre) et de 15,7 % (au nord) par rapport à la sous-période d’avant. Ce qui n’est pas sans conséquences sur la dynamique hydrologique du secteur d’étude. Ainsi, il est observé une régression de 49,27 % des débits moyens annuels entre les sous-périodes 1951-1968 et 1969-1987 et une augmentation de 65 % entre les sous-périodes 1969-1987 et 1988-2015. Sur la dernière sous-période (1988-2015), les résultats révèlent aussi une augmentation des indicateurs des évènements hydroclimatiques extrêmes susceptibles d’occasionner les inondations à Bonou. Les pluies maximales journalières ont une occurrence de 2 ans, 5 ans, 10 ans pour les événements pluvieux forts ; 20 ans, pour les événements pluvieux très forts ; 50 ans et + pour les événements pluvieux extrêmement forts dans le secteur d’étude. Les débits maximaux en 24 heures augmentent en même temps que les périodes de retour. Les débits maximaux de 50ème quantile surviennent chaque deux ans, les débits maximaux correspondants aux évènements hydrologiquement forts, très forts, surviennent tous les 10 ans et les débits extrêmement forts ont une occurrence centennale. Les conditions atmosphériques, océaniques, de la surface continentale, la vitesse du vent, l’humidité relative, les facteurs géographiques sont autant de facteurs qui interagissent dans la répartition spatio-temporelle de la pluie du secteur d’étude. La pluviométrie du bassin est également influencée par le flux de mousson (nord secteur d’étude) auquel s’associent les lignes de grains de sud-est, responsables des pluies orageuses et d’averses qui génèrent des inondations dans le bassin. De même, le contexte climatique actuel et la dynamique de l’occupation des terres constituent des déterminants qui amplifient la manifestation des inondations dans le secteur d’étude. La recherche indique que le modèle conceptuel pluie-débit GR4J, surestime les débits observés en période de basse eau et les sous-estime en période de hautes eaux. Les critères d’efficacité et de performance NSE, RMSE et KGE, mis en évidence et calculés sur les débits de hautes eaux ont donné des résultats meilleurs en calage qu’en validation. Mieux les valeurs du KGE varient de 83 à 85 % en calage et de 56 à 68 % en validation, ce qui confère au modèle GR4J, l’efficacité, la performance à reproduire les débits extrêmes de crues dans le bassin. Le GR4J peut donc être servir comme outil d’aide à la décision pour l’actualisation des normes hydrologiques dans le secteur d’étude. Dans la perspective de mieux étudier les évènements climatiques extrêmes futurs, cette recherche a évalué des données de projection de sorties des modèles climatiques régionaux issues du Programme CORDEX africain et qui pourraient être utilisées pour les projections hydrologiques des travaux futurs dans le secteur d’étude. / This research aims to characterize the indicators relating to extreme hydroclimatic events in Ouémé Watershed at Bonou’s outlet. To achieve this objective, climatological data (daily rain, daily temperatures, potential daily evapotranspiration), hydrometric data (daily flows), reanalysis data (atmospheric pressure, relative humidity, wind speed, West African Monsoon Indices), the ocean surface temperature on the Cotonou coast and land use mapping data of 1986, 2000, 2015 are used. The analysis of the hydroclimatic variability revealed that during the 1951-2015 period, the study area is marked by two breaks stationarity (1968 and 1987) which divided the series into three sub-periods. The rainfall variability reveals a regression of 18 % (in the south), 16 % (in the center) and 15.8 % (in the north) between the first sub-period (1951-1968) and the second sub-period (1969-1987), while the last sub-period (1988-2015) is marked by a rainfall recovery of 15 % (in the south), 16.3 % (in the center) and 15.7 % (in north) relative to the previous sub-period (1969-1987). This is not without consequences on the hydrological dynamics of the study area. In this sense, it is noted that there is a regression of 49.27% of the average annual flows between the 1951-1968 and 1969-1987 sub-periods and a 65 % increase between the 1969-1987 and the 1969-1987 sub-periods. 1988-2015. In the last sub-period (1988-2015), the results also reveal an increase in the indicators of extreme hydroclimatic events likely to cause floods Ouémé Watershed at Bonou’s outlet. The results shows that daily maximum rainfall has an occurrence of 2 years, 5 years, 10 years for strong rainfall events; 20 years, for very heavy rain events, 50 years and over for extremely heavy rain events in the study area. The maximum flow rates in 24 hours at the sametime as the return periods. The maximum flow rates of 50th quantiles occur every two years (02 years), the maximum flows corresponding to the hydrologically strong events, very strong occur every 10 years and the extremely high flows have an occurrence centennial. The conditions of atmospheric, oceanic, continental surface, wind speed, relative humidity, and geographic factors are all elements that interact in the spatio-temporal distribution of rainfall in the study area. The Rainfall of the basin is also influenced by the monsoon flow (north of the study area), to which the southeastern grain lines are associated and are responsible for the stormy rains and showers that generate flooding in the area. In addition, the current climatic context and land-use dynamics are determinants that amplify the occurrence of floods in the study area. The reasearch indicates that the rain-flow conceptual model, of the Rural Engineering with four parameters and the daily time step (GR4J), overestimates the flows observed during the low water period and the sub-peaks. The efficiency and performance criteria NSE, RMSE and KGE, highlighted and calculated on high water flow rates, gave better results in calibration than in validation. Better KGE values range from 83 to 85 % calibration and from 56 to 68 % validation, which gives the GR4J model the efficiency and performance to reproduce extreme flows of floods in the study area. The GR4J can therefore be used as a decision to support tool for updating hydrological standards in the study area. In order to future study about extreme climatic events, this research assessed output projection data from regional climate models from the African CORDEX Program that could be used for hydrological projections in future work in Ouémé Watershed at Bonou’s outlet.

Pendant plus de six décennies, les antivitamines K (AVK) ont été la seule classe d’anticoagulants oraux disponibles sur le marché. L’arrivée des Anticoagulants Oraux Directs (AOD) en 2008 a marqué un réel tournant dans l’anticoagulation par voie orale. La première partie de ce travail s’est intéressée aux caractéristiques des patients traités par anticoagulants, et les conséquences en termes d’hospitalisations pour évènement hémorragique ou ischémique. Au sein d’une cohorte de patients hospitalisés pour AVC (cohorte BIOSTROKE, Lille), aucun facteur d’influence particulier, intervenant sur la survenue et l’évolution (mortalité, de déclin cognitif, et de handicap à 3 mois) de l’AVC n’était mis en évidence. En parallèle, une étude menée dans le service d’urgences du Centre Hospitalier de Béthune sur 3 années (2012, 2014, 2016) pour suivre l’évolution des prescriptions d’anticoagulants oraux et observer l’impact de la mise sur le marché des AOD, montrait une population dont certaines caractéristiques semblaient être associées à une utilisation plus volontiers des AOD ou des AVK. Le nombre d’évènements hémorragiques et thrombotiques sous AVK restait cependant stable sur les trois périodes d’étude (770 patients inclus au total). Forts de ces données rassurantes, la deuxième partie de ce travail s’est intéressée aux pratiques de prescription et au bon usage des anticoagulants, avec un focus particulier sur les AOD, à travers trois études : la première réalisée auprès des médecins généralistes du Nord et du Pas de Calais qui montrait une population de prescripteurs plutôt méfiants envers les AOD et ayant une préférence pour les anti-Xa. Les deux autres études portaient sur le bon usage des AOD, à partir de prescriptions à l’officine (grâce à la participation des étudiants en pharmacie de la faculté de Lille en stage d’officine) et à l’hôpital (Centre Hospitalier Régional de Lille). Ces études retrouvaient globalement des chiffres de prescriptions d’AOD transposables aux données françaises, avec cependant des disparités ville/hôpital. En termes de bon usage, les prescriptions étaient pour moins d’un tiers d’entre elles non-conformes aux recommandations en vigueur, avec une forte proportion de situations de sous-dosages. En conclusion, ce travail de thèse a permis de mettre à jour des pistes de formation/information des professionnels de santé sur ces nouvelles molécules anticoagulantes qui vont participer à une prise en charge avisée et optimale des patients. / For over six decades, vitamin K antagonists (VKAs) were the only class of oral anticoagulants available on the market. The arrival of Direct Oral Anticoagulants (DOAs) in 2008 marked a real turning point in the oral anticoagulation. The first part of this work was interested in the characteristics of patients treated with anticoagulants and the consequences in terms of hospitalizations for ischemic or hemorrhagic events. In a cohort of patients hospitalized for stroke (cohort BIOSTROKE Lille), no particular influencer, on the onset and course (mortality, cognitive decline, and disability at 3 months) of stroke was highlighted. In parallel, a study conducted in the emergency department of the Hospital of Bethune on three years (2012, 2014, 2016) to observe oral anticoagulants prescriptions and the impact of DOAs arriving, showed a population where many characteristics appear to be associated with a more readily using DOAs or VKAs. The number of bleeding and thrombotic events with VKAs remained stable during the three study periods (in total 770 patients were included). With these reassuring data, the second part of this work has focused on prescribing practices and the appropriate use of anticoagulants, with a particular focus on DOAs, through three studies: one conducted among general practitioners of Nord and Pas de Calais departments, which showed a rather wary about DOAs but with a preference for anti-Xa. The other two studies focused on the appropriate use of DOAs, estimated from pharmacy (with the participation of pharmacy students of the Faculty of Lille) and hospital (Lille Regional Hospital Center) prescriptions. Data about DOAs prescriptions were transposable to French data, albeit with city / hospital disparities. Non-appropriate prescriptions were observed in less than one-third file, with a high proportion of under dosing situations. In conclusion, this work has allowed updating training tracks / information for health care professionals on these new anticoagulant molecules to improve information and optimal patients’ care.

We study interdependencies in spousal labor supply and the effectiveness of intrahousehold insurance in a sample of married couples, where the husband loses his job due to a mass layoff or plant closure using data from the Austrian Social Security Database. We show that in our sample of relatively young couples the shock hits households at crucial stages of family formation, which requires careful modeling of the wives' counterfactual lifecycle labor market patterns. In our empirical analysis, we propose three independent control groups of unaffected households to identify the causal effects of husbands' displacement on wives' labor supply. Our empirical results show that husbands suffer large and persistent employment and earnings losses over the first 5 years after displacement. But wives' labor supply increases only moderately and they respond predominantly at the extensive margin. The implied participation elasticity with respect to the husband's earnings shock is very small, about -0:04. While the wives' earnings gains recover only a tiny fraction of the household income loss, public transfers and taxes are a more important insurance at least in the short run. In terms of non-labor market related outcomes, we find a small positive effect on the probability of divorce, but no effect of the husband's job displacement on fertility. The presence and ages of children in the household are crucial determinants of the wife's labor supply response. The most responsive group are mothers, who are planning to return to the labor market after a maternity break, while mothers of very young children or wives without children remain unresponsive. We thus conclude that Austria's strong gender identity norms are an explanation for the limited scope of intra-household insurance.

Each of us lives in some municipality or at least in the cadastral district of some municipality. The council of each municipality deals daily with many issues connected with routine and visible operation of municipality. Unfortunately, sometimes it is necessary to be prepared to solve also the exceptional events or even crisis situations. The author aims to create the summarization which will resume the duties of mayors connected with crisis management and which will help the mayors with long examination of this issue. This thesis covers the Pilsen region because this is the author's home town. To create such a document, it is necessary to perform following operations. First of all, it is required to cover the whole theory of this issue which is laid down in several legislative documents collectively named "the Crisis Legislation". Processing of these legislative documents is performed to correspond to topic of this thesis and not to include the useless information beyond the frame of the topic. The name of the thesis is "Obligations of mayors at solution of emergency and crisis situations in Pilsen region". Indivisible part of dealing with exceptional event and crisis situation is also the preparation for its´ solving within period before occurrence of the exceptional event or crisis situation. This is also covered in processing of the legislative documents. The research part of this thesis focuses on knowledge of mayors to clarify the real need to create this summarization and also the required particularity of this document. This research was conducted by means of questionnaires containing fifteen questions. Five of them are YES/NO questions and the left ten questions are close-ended with one correct answer. The YES/NO (so called informative) questions and answers indicate whether the questionnaires were filled out by the mayors of either cities or villages, what are the sources of threat and several other information. The electronic form was filled out by 100 mayors of the total 501 municipalities of the Pilsen Region. To have the possibility to compare the results of mayors with other results, the same questionnaire in the paper form with only 10 close-ended questions with one correct answer was sent also to 50 laymen from the social environment of the author. On basis of these researches, following hypotheses were pronounced: H1 Mayors and chairmen in the Pilsen Region have the knowledge distributed close to Poisson distribution. H2 Laymen from social environment of the author have knowledge distributed close to normal distribution. H3 Mayors in the Pilsen Region have higher level of knowledge than the laymen from the social environment of the author. The H1 and H2 hypotheses were verified by the 2 test on the level of statistic relevance =0,05, whereas the H1 was refuted and the H2 was confirmed. The H3 hypothesis was then confirmed by the double-choice T-test. Each question of the questionnaire (inclusive the informative questions for mayors) was individually visualised in the diagram, which presents the percentual fruitfulness of mayors and of laymen. These diagrams show obviously that the majority of mayors who filled up the questionnaire, are the mayors of smaller municipalities. More than a half (63%) of these municipalities have less than 500 inhabitants. Large part of these mayors are probably mayors with a full-time employment. The results of mayors did not confirm the distribution of knowledge close to Poisson distribution and the total fruitfulness of mayors´ answers was 61%. This signifies that it is desirable to create this summarization rather as more detailed document with concentration on smaller municipalities and as a chart. The author believes that this summarization will be useful for the mayors.

The diploma thesis deals with design and implementation of cooperative work on documents in the application TEXonWeb. The diploma theses includes the analysis of competitive solutions and key technologies for implementation. The cooperative work on documents is a functionality which allows users to cooperate simply on one document in real-time. Text of the thesis describes the process of implementation using the technology which is described in the initial analysis. The diploma theses also includes some ideas for next possible development.