rio: Dengue Fever in the State of Rio de Janeiro
In rbugs: Fusing R and OpenBugs and Beyond

Description Usage Format Examples

Shape file of the state of Rio de Janeiro

data(rio)

A shape file with 92 observations on the following 9 variables.

GeoCode: Rio de Janeiro counties numeric codes
Name: Rio de Janeiro counties names
Lat: Rio de Janeiro counties latitude
Long: Rio de Janeiro counties longitude
Pop: Rio de Janeiro counties population
Income: Rio de Janeiro counties percentage of homes living with less than one minimun salary
Urban: Rio de Janeiro counties percentage of urban area
Dengue: Counts of dengue fever cases by county in the state of Rio de Janeiro
E: Expected counts of dengue fever cases by county in the state of Rio de Janeiro, conditioned on the total number of cases the stratified by population rate

data(rio)

## Not run: 
library(spdep)
rio.bug <- system.file("bugs/model/dengue.bug", package="rbugs")

## Get the neighbor structure for Rio
rj.nb=poly2nb(rio)
 
## Number of cases by county
Y = rio@data$Dengue

## Get covariates
X <- as.matrix(rio@data[,c("Urban","Income")])

## Get expected values
E <- rio@data$E

## Number of covariates
P <- ncol(X)

## Total counties
N <- length(rj.nb)
 
## Number of neighbors of each county
num <- sapply(rj.nb, length)
 
## Adjacency neighbor list of each county
adj <- unlist(rj.nb)

## Total sum of the number of neighbors in the map
sumNumNeigh <- length(adj)
 
## Set data file
data.rio <- list (N=N, P=P, Y=Y, adj=adj, X=X, E=E, num=num, sumNumNeigh=sumNumNeigh)

## Generate the intial values for spatial vector (u), random noise (v) and covariates effects (beta)
u.aux <- rep(0,N)
v.aux <- rep(0,N)
beta.aux <- rep(0,P)

## Generate the list with the nescessary information in the init file
inits <- list( list(beta = beta.aux, tau.u = 1, tau.v = 1, u=u.aux, v=v.aux, Int=0.0))

## Set the parameters that will be saved and returned by BUGS
parameters <- c("beta", "tau.u", "tau.v", "RR", "Int")

## no tested examples for mac-os.

rio.sim <- rbugs(data.rio, inits, parameters,
                 rio.bug, n.chains=2, n.iter=60000, 
                 n.burnin = 10000, n.thin = 10, 
                 bugs="/usr/bin/OpenBUGS",
                 bugsWorkingDir="/home/marcos/tmp")

## MCMC analysis
library("coda")
rio.mcmc <- rbugs2coda(rio.sim)
summary(rio.mcmc)
effectiveSize(rio.mcmc)
gelman.diag(rio.mcmc)



## End(Not run)