Linking Wildbook and marked via RWildbook"
In RWildbook: Interface for the 'Wildbook' Wildlife Data Management Framework

Here we show how data retrieved from a Wildbook database can be formatted and analysed using the marked package by Jeff Laake (also the creator of the RMark package). In particular, we fit a simple Cormack-Jolly-Seber model to data retrieved from whaleshark.org.

First we load the RWildbook and marked package:

## Load other packages
library(RWildbook)
library(marked)

Next we use the searchWB() function from RWildbook to retrieve data for the first 99 whale sharks:

## Extract data for individual A-001 through A-099
vignette_2_data <- searchWB(username="username",
                            password="password",
                            baseURL ="whaleshark.org",
                            object="Encounter",
                            individualID=paste0("A-0",rep(0:9,rep(10,10)),rep(0:9,10))[-1])

This results in the following data frame (not the that the time of encounter has been modified to maintain data privacy):

data("vignette_2_data",package="RWildbook")
head(vignette_2_data)[,c("individualID","location","year","month","day")]

Now we can create capture histories by defining the start and end times of the capture occasions. For illustration, we assume that the capture occasions cover January, February, and March of each year from 1998 through 2016.

 ## Define start and end dates of capture occasions
 start.dates1 <- paste0(1998:2016,"-01-01") #Define the start.date value
 end.dates1 <- paste0(1998:2016,"-04-01") #Define the end.date value

 ## Format data for use in marked
 markedData1 <- markedData(data = vignette_2_data,
                              varname_of_capturetime = "dateInMilliseconds",
                              varlist = c("individualID"),
                              start.dates = start.dates1,
                              end.dates = NULL,
                              date_format = "%Y-%m-%d",
                              origin = "1970-01-01")

We can now fit the Cormack-Jolly-Seber model using the functions in mark to first format the data:

 ## Fit simple CJS model in marked
 markedData1.proc=process.data(markedData1,model="cjs",begin.time=1)
 markedData1.ddl=make.design.data(markedData1.proc)
 markedData1.cjs=crm(markedData1.proc,markedData1.ddl,model.parameters=list(Phi=list(formula=~time),p=list(formula=~time)))

and plot the estimated capture and survival probabilities:

## Plot parameter estimates
plot(1998:2015,markedData1.cjs$results$reals$Phi$estimate,
        pch=16,col="green",
        main="Estimated Capture and Survival Probabilities",
        xlab="Occasion",ylab="Estimate",ylim=c(0,1))
points(1999:2016,markedData1.cjs$results$reals$p$estimate,pch=16,col="blue")
legend("bottomright",pch=16,col=c("green","blue"),legend=c("Survival","Capture"))