Geospatial Science and Technology: Projects
Models of Infectious Disease Agent Study (MIDAS)
MIDAS is a collaborative network of scientists who lead research in the use of computational and mathematical models (with an emphasis on agent-based models) that will prepare the nation to respond to outbreaks of infectious disease. MIDAS's research mission includes investigations of the dynamics of emergence and spread of pathogens and their products; identification and surveillance of infectious diseases; effectiveness and consequence of intervention strategies; host/pathogen interactions; and ecological, climatic, and evolutionary dimensions of infectious disease outbreaks.
As the core informatics group providing computational cycles, data support, and computer science research, we support the MIDAS mission providing LINUX cluster computation resources, large-scale geospatial database design and development, and a variety of research support including statistical analysis of historical past pandemics to provide modelers with data, model parameters, and computational resources. Geospatial data is a key component of the work of MIDAS, and we have developed large-scale geospatial data resources to support modeling.
Key among the innovative work done on MIDAS has been the development of a synthesized micro-level national human population database. In the parlance of agent-based modeling, these micro-level data on individuals represent the 'agents' that are modeled. In this spatially explicit database, every individual and family living in a household is represented. These households and families are tied to realistic household locations based on census block group data. The resulting database provides a national data infrastructure for infectious disease modelers. Because schools and workplaces are important places where social contact can increase disease transmission, we also developed innovative methods for assigning these synthesized agents to schools and to workplaces (based on age characteristics and national level data on school locations and sizes and national level data on workplaces and commuting patterns). Modelers can now explicitly model disease transmission within families, at schools, and at workplaces.
A second key innovation was the development of a synthesize poultry farm geospatial database. Using many national geospatial databases and counts of poultry farms (by size and type by county), we developed a suitability model and located each 'synthesized' farm in a way that reproduces the like distribution of poultry farms in the U.S. Having these synthesized, realistic poultry farm locations along with correct distributions of farm type and farm size provides modelers with a national dataset on poultry farms that did not previously exist. Veterinarian modelers use these data to study disease transmission and potential mitigation strategies in the commercial poultry industry.
rti.org