Stochastic microsimulation models require data to populate both probability distributions for events/outcomes and create exact values for characteristics of individuals in the model (e.g. gender, age). This presentation is based on an agent-based model, still in development, that is designed to predict salmonellosis rates in humans, where the origin is from eggs. Multiple data sources are required to construct this model, two of which are official data sources. Census data is being used to construct individuals, using a synthetic population method. Salmonellosis is a notifiable disease in New Zealand, and annual salmonellosis rates are published by Environmental Science & Research, a Crown Research Institute in New Zealand. Data is also being obtained from government data sources that are not official. Ministry of Health reports have been used to estimate the contamination rates of chicken eggs in New Zealand. Food Standards Australia New Zealand conducted a 2008 survey on egg consumption and behavioural factors associated with egg-related food-borne illness. Finally, data from non-government sources is also being used. For person-to-person transmission, data on bathroom handwashing rates is being sourced from the peer reviewed literature. Salmonellosis duration has been estimated based on data also from the peer-reviewed literature. This presentation will describe the logic behind identifying the data needed for the simulation. Often, the data is not in a suitable form for use in a microsimulation, and this presentation will outline the approaches used to change the data into the required inputs. Finally, a description of how all the data sources are being integrated into the microsimulation will be outlined. This microsimulation forms part of a PhD project and comments from the audience are welcomed.