Sunbelt XXXI International Network for Social Network ... - INSNA

Finding A Cure For Crohn’s Disease ‐ Analyzing Medical Conditions In Online CommunitiesFuehres, Hauke; Gloor, Peter A.; Kaminski, Jermain; Zhang, XueWords and Networks ‐ Roles, Health, MethodsNetworks And Health, Chronic Disease, Drug Use, Communities Of Practice, Semantic Networks, Data MiningFRI.PM1We introduce a project trying to automatically collect information from patients of a chronic disease by analyzing online social networks. More specifically, welook at online postings of patients with Crohn’s Disease and Ulcerative Colitis by analyzing their Facebook friendship and discussion networks as well as Twitterand other online forum posts. We identified Facebook groups whose goal is to improve the lives of Crohn’s patients, or even to find a cure. This informationprovides valuable input to the creation process of creating a community of Crohn’s patients. The friendship network of the Facebook groups is surprisinglyunconnected, even considering that we are only able to collect the links of people who have made their friends public. There are only little star networks withvery short average path length of three to four. When we manually checked a few of them, the person in the center was usually somebody affected withCrohn’s disease, and the people in the periphery where their friends. This does not mean that the people in the center don’t have bridge links through friendof‐friendconnections. Rather, this means that the fans of the Crohn’s support pages don’t know each other. Additionally we also did a content analysis ofFacebook and Twitter, which illustrates very different medical approaches in different regions, with emphasis on specific drugs like Prednisone in the US, ondiets and other drugs like Humira in the UK, and Remicade and surgery in Canada.Finding A Needle In A Haystack: Advances And Challenges In Using Blau Space Modeling To Trace Covert Social NetworksBrashears, Matthew E.; Genkin, MichaelCriminals, Gangs, Terrorists, and NetworksTheory, Terrorism, Criminal Behavior, Covert, Blau SpaceSAT.PM1How can covert social networks be detected and mapped? Such networks, including terror cells, drug cartels, and human smuggling rings are of considerableinterest to scholars but are very difficult to study. We discuss a three‐year project funded by the Defense Threat Reduction Agency to develop methods foridentifying such covert social networks. We will identify some of the theoretical challenges involved in detecting covert social networks and discuss the BlauSpace model that is the foundation for our effort. Specifically, we hope to use network features such as bridge signatures and pendant status, in conjunctionwith Blau Space location to provide reliable indications of involvement in covert groups. We additionally present some preliminary analyses using the NationalLongitudinal Study of Adolescent Health (Add Health). This dataset is particularly useful for method‐validation because it is highly detailed in respect to BlauSpace parameters and contains information on deviant networks, thereby allowing us to check whether the deviant ties we predict correspond to actualdeviant ties. While the adolescents included in this dataset are unlikely to be involved in serious covert activities, they represent a useful test case for thedevelopment of our techniques. Directions in extending our model theoretically and to other data sources will also be discussed.