IFIMAR   20926
INSTITUTO DE INVESTIGACIONES FISICAS DE MAR DEL PLATA
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
Creating “real social networks” from Census Data
Autor/es:
M. V. MIGUELES; C. LAGORIO; P. A. MACRI; L. A. BRAUNSTEIN
Lugar:
Pinamar, Provincia de Buenos Aires
Reunión:
Workshop; DYSES; 2009
Resumen:
We study the behavior of a society when it is exposed to different socio-economic phenomena.This study is hard to perform in a real society mainly because the data is unavailable and alsobecause it is hard to manipulate so many variables in order to create different scenarios. In orderto recreate a “real society”, models, simulations and visualizations are powerful tools. Networktheory has proven to be very successful to model social interactions. In these networks the nodesare the individuals and the contact between them their interaction. These “social networks” represents very well the heterogeneity of the population. In our work, we create a bipartite network (a network with two types of nodes), the first type of node represents the people and the other represents the locations where people meet. In order to create people-nodes we use a synthetic population algorithm call Iterational Proportional Fitting Procedure (IPFP) , which uses aggregated and disaggregated data from the Census to create a synthetic population with the same statistical properties than the real population. Some properties of the location-nodes are also provided by Census data. These nodes have also geographic positions that correspond to a place in the real city. Departingfrom the bipartite network we can establish different contacts networks, defining which type of contact (links) is relevant for the measurement we want to accomplish. Using a visual interface we can control all the variables of synthetic population and contact places. We apply this program to epidemics models where the social networks as the underlying substrate were the transmission take place. We define a contact (link) when two people share the same contact place. Using this network as underlying substrate, we performed several measurements such as the time evolution of infected individuals, time distribution of the duration of an epidemic, final size of the epidemics, etc .
rds']