Learn R Programming

FSA (version 0.9.5)

capHistConvert: Convert between capture history data.frame formats.

Description

Use to convert between simple versions of several capture history data.frame formats -- “individual”, “frequency”, “event”, “MARK”, and “RMark”. The primary use is to convert to the “individual” format for use in capHistSum.

Usage

capHistConvert(
  df,
  cols2use = NULL,
  cols2ignore = NULL,
  in.type = c("frequency", "event", "individual", "MARK", "marked", "RMark"),
  out.type = c("individual", "event", "frequency", "MARK", "marked", "RMark"),
  id = NULL,
  event.ord = NULL,
  freq = NULL,
  var.lbls = NULL,
  var.lbls.pre = "event",
  include.id = ifelse(is.null(id), FALSE, TRUE)
)

Value

A data frame of the proper type given in out.type is returned. See details.

Arguments

df

A data.frame that contains the capture histories and, perhaps, a unique fish identifier or frequency variable. See details.

cols2use

A string or numeric vector that indicates columns in df to use. Negative numeric values will not use those columns. Cannot use both cols2use and col2ignore.

cols2ignore

A string or numeric vector that indicates columns in df to ignore. Typical columns to ignore are those that are not either in id= or freq= or part of the capture history data. Cannot use both cols2use and col2ignore.

in.type

A single string that indicates the type of capture history format to convert FROM.

out.type

A single string that indicates the type of capture history format to convert TO.

id

A string or numeric that indicates the column in df that contains the unique identifier for an individual fish. This argument is only used if in.type="event", in.type="individual", or, possibly, in.type="RMark".

event.ord

A string that contains a vector of ordered levels to be used when in.type="event". The default is to order alphabetically which may not be desirable if, for example, the events are labeled as ‘first’, ‘second’, ‘third’, and ‘fourth’. In this case, use event.ord=c("first","second","third","fourth").

freq

A string or numeric that indicates the column in df that contains the frequency of individual fish corresponding to a capture history. This argument is only used if in.type="MARK", in.type="frequency", or, possibly, in.type="RMark".

var.lbls

A string vector of labels for the columns that contain the returned individual or frequency capture histories. If var.lbls=NULL or the length is different then the number of events then default labels using var.lbls.pre will be used. This argument is only used if out.type="frequency" or out.type="individual".

var.lbls.pre

A single string used as a prefix for the labels of the columns that contain the returned individual or frequency capture histories. This prefix will be appended with a number corresponding to the sample event. This argument is only used if out.type="frequency" or out.type="individual" and will be ignored if a proper vector is given in var.lbls.

include.id

A logical that indicates whether a unique fish identifier variable/column should be included in the output data.frame. This argument is only used if out.type="individual" or out.type="RMark".

Warning

capHistConvert may give unwanted results if the data are in.type="event" but there are unused levels for the variable, as would result if the data.frame had been subsetted on the event variable. The unwanted results can be corrected by using droplevels before capHistConvert. See the last example for an example.

IFAR Chapter

9-Abundance from Capture-Recapture Data.

Author

Derek H. Ogle, DerekOgle51@gmail.com

Details

capHistSum requires capture histories to be recorded in the “individual” format. In this format, the data frame contains (at least) as many columns as sample events and as many rows as individually tagged fish. Optionally, the data.frame may also contain a column with unique fish identifiers (e.g., tag numbers). Each cell in the capture history portion of the data.frame contains a ‘0’ if the fish of that row was NOT seen in the event of that column and a ‘1’ if the fish of that row WAS seen in the event of that column. For example, suppose that five fish were marked on four sampling events; fish ‘17’ was captured on the first two events; fish ‘18’ was captured on the first and third events; fish ‘19’ was captured on only the third event; fish ‘20’ was captured on only the fourth event; and fish ‘21’ was captured on the first and second events. The “individual” capture history date.frame for these data looks like:

fishevent1event2event3event4
171100
181010
190010
200001
211100

The “frequency” format data.frame (this format is used in Rcapture) has unique capture histories in separate columns, as in the “individual” format, but also includes a column with the frequency of individuals that had the capture history of that row. It will not contain a fish identifier variable. The same data from above looks like:

event1event2event3event4freq
11002
10101
00101
00011

The “event” format data.frame has a column with the unique fish identifier and a column with the event in which the fish of that row was observed. The same data from above looks like:

fishevent
171
181
211
172
212
183
193
204

MARK (http://www.phidot.org/software/mark/index.html) is the “gold-standard” software for analyzing complex capture history information. In the “MARK” format the 0s and 1s of the capture histories are combined together as a string without any spaces. Thus, the “MARK” format has the capture history strings in one column with an additional column that contains the frequency of individuals that exhibited the capture history of that row. The final column ends with a semi-colon. The same data from above looks like:

chfreq
00011;
00101;
10101;
11002;

The RMark and marked are packages used to replace some of the functionality of MARK or to interact with MARK. The “RMark” or “marked” format requires the capture histories as one string (must be a character string and called ‘ch’), as in the “MARK” format, but without the semicolon. The data.frame may be augmented with an identifier for individual fish OR with a frequency variable. If augmented with a unique fish identification variable then the same data from above looks like:

fishch
171100
181010
190010
200001
211100

However, if augmented with a frequency variable then the same data from above looks like:

chfreq
00011
00101
10101
11002

Each of the formats can be used to convert from (i.e., in in.type=) or to convert to (i.e., in out.type=) with the exception that only the individual fish identifier version can be converted to when out.type="RMark".

References

Ogle, D.H. 2016. Introductory Fisheries Analyses with R. Chapman & Hall/CRC, Boca Raton, FL.

See Also

See capHistSum to summarize “individual” capture histories into a format usable in mrClosed and mrOpen. Also see Rcapture, RMark, or marked packages for handling more complex analyses.

Examples

Run this code
## A small example of 'event' format
( ex1 <- data.frame(fish=c(17,18,21,17,21,18,19,20),yr=c(1987,1987,1987,1988,1988,1989,1989,1990)) )
# convert to 'individual' format
( ex1.E2I <- capHistConvert(ex1,id="fish",in.type="event") )
# convert to 'frequency' format
( ex1.E2F <- capHistConvert(ex1,id="fish",in.type="event",out.type="frequency") )
# convert to 'MARK' format
( ex1.E2M <- capHistConvert(ex1,id="fish",in.type="event",out.type="MARK") )
# convert to 'RMark' format
( ex1.E2R <- capHistConvert(ex1,id="fish",in.type="event",out.type="RMark") )

## convert converted 'individual' format ...
# to 'frequency' format (must ignore "id")
( ex1.I2F <- capHistConvert(ex1.E2I,id="fish",in.type="individual",out.type="frequency") )
# to 'MARK' format
( ex1.I2M <- capHistConvert(ex1.E2I,id="fish",in.type="individual",out.type="MARK") )
# to 'RMark' format
( ex1.I2R <- capHistConvert(ex1.E2I,id="fish",in.type="individual",out.type="RMark") )
# to 'event' format
( ex1.I2E <- capHistConvert(ex1.E2I,id="fish",in.type="individual",out.type="event") )

#' ## convert converted 'frequency' format ...
# to 'individual' format
( ex1.F2I <- capHistConvert(ex1.E2F,freq="freq",in.type="frequency") )
( ex1.F2Ia <- capHistConvert(ex1.E2F,freq="freq",in.type="frequency",include.id=TRUE) )
# to 'Mark' format
( ex1.F2M <- capHistConvert(ex1.E2F,freq="freq",in.type="frequency",
                            out.type="MARK") )
# to 'RMark' format
( ex1.F2R <- capHistConvert(ex1.E2F,freq="freq",in.type="frequency",
                            out.type="RMark") )
( ex1.F2Ra <- capHistConvert(ex1.E2F,freq="freq",in.type="frequency",
                             out.type="RMark",include.id=TRUE) )
# to 'event' format
( ex1.F2E <- capHistConvert(ex1.E2F,freq="freq",in.type="frequency",
                            out.type="event") )

## convert converted 'MARK' format ...
# to 'individual' format
( ex1.M2I <- capHistConvert(ex1.E2M,freq="freq",in.type="MARK") )
( ex1.M2Ia <- capHistConvert(ex1.E2M,freq="freq",in.type="MARK",include.id=TRUE) )
# to 'frequency' format
( ex1.M2F <- capHistConvert(ex1.E2M,freq="freq",in.type="MARK",out.type="frequency") )
# to 'RMark' format
( ex1.M2R <- capHistConvert(ex1.E2M,freq="freq",in.type="MARK",out.type="RMark") )
( ex1.M2Ra <- capHistConvert(ex1.E2M,freq="freq",in.type="MARK",out.type="RMark",include.id=TRUE) )
# to 'event' format
( ex1.M2E <- capHistConvert(ex1.E2M,freq="freq",in.type="MARK",out.type="event") )
 
## convert converted 'RMark' format ...
# to 'individual' format
( ex1.R2I <- capHistConvert(ex1.E2R,id="fish",in.type="RMark") )
# to 'frequency' format
( ex1.R2F <- capHistConvert(ex1.E2R,id="fish",in.type="RMark",out.type="frequency") )
# to 'MARK' format
( ex1.R2M <- capHistConvert(ex1.E2R,id="fish",in.type="RMark",out.type="MARK") )
# to 'event' format
( ex1.R2E <- capHistConvert(ex1.E2R,id="fish",in.type="RMark",out.type="event") )

## Remove semi-colon from MARK format to make a RMark 'frequency' format
ex1.E2R1 <- ex1.E2M
ex1.E2R1$freq <- as.numeric(sub(";","",ex1.E2R1$freq))
ex1.E2R1
# convert this to 'individual' format
( ex1.R2I1 <- capHistConvert(ex1.E2R1,freq="freq",in.type="RMark") )
( ex1.R2I1a <- capHistConvert(ex1.E2R1,freq="freq",in.type="RMark",include.id=TRUE) )
# convert this to 'frequency' format
( ex1.R2F1 <- capHistConvert(ex1.E2R1,freq="freq",in.type="RMark",out.type="frequency") )
# convert this to 'MARK' format
( ex1.R2M1 <- capHistConvert(ex1.E2R1,freq="freq",in.type="RMark",out.type="MARK") )
# convert this to 'event' format
( ex1.R2E1 <- capHistConvert(ex1.E2R1,freq="freq",in.type="RMark",out.type="event") )


########################################################################
## A small example using character ids
( ex2 <- data.frame(fish=c("id17","id18","id21","id17","id21","id18","id19","id20"),
                    yr=c(1987,1987,1987,1988,1988,1989,1989,1990)) )
# convert to 'individual' format
( ex2.E2I <- capHistConvert(ex2,id="fish",in.type="event") )
# convert to 'frequency' format
( ex2.E2F <- capHistConvert(ex2,id="fish",in.type="event",out.type="frequency") )
# convert to 'MARK' format
( ex2.E2M <- capHistConvert(ex2,id="fish",in.type="event",out.type="MARK") )
# convert to 'RMark' format
( ex2.E2R <- capHistConvert(ex2,id="fish",in.type="event",out.type="RMark") )

## convert converted 'individual' format ...
# to 'frequency' format
( ex2.I2F <- capHistConvert(ex2.E2I,id="fish",in.type="individual",out.type="frequency") )
# to 'MARK' format
( ex2.I2M <- capHistConvert(ex2.E2I,id="fish",in.type="individual",out.type="MARK") )
# to 'RMark' format
( ex2.I2R <- capHistConvert(ex2.E2I,id="fish",in.type="individual",out.type="RMark") )
# to 'event' format
( ex2.I2E <- capHistConvert(ex2.E2I,id="fish",in.type="individual",out.type="event") )

## demo use of var.lbls
( ex2.E2Ia <- capHistConvert(ex2,id="fish",in.type="event",var.lbls.pre="Sample") )
( ex2.E2Ib <- capHistConvert(ex2,id="fish",in.type="event",
              var.lbls=c("first","second","third","fourth")) )

## demo use of event.ord
( ex2.I2Ea <- capHistConvert(ex2.E2Ib,id="fish",in.type="individual",out.type="event") )
( ex2.E2Ibad <- capHistConvert(ex2.I2Ea,id="fish",in.type="event") )
( ex2.E2Igood <- capHistConvert(ex2.I2Ea,id="fish",in.type="event",
                 event.ord=c("first","second","third","fourth")) )

## ONLY RUN IN INTERACTIVE MODE
if (interactive()) {

########################################################################
## A larger example of 'frequency' format (data from Rcapture package)
data(bunting,package="Rcapture")
head(bunting)
# convert to 'individual' format
bun.F2I <- capHistConvert(bunting,in.type="frequency",freq="freq")
head(bun.F2I)
# convert to 'MARK' format
bun.F2M <- capHistConvert(bunting,id="id",in.type="frequency",freq="freq",out.type="MARK")
head(bun.F2M)
# convert converted 'individual' back to 'MARK' format
bun.I2M <- capHistConvert(bun.F2I,id="id",in.type="individual",out.type="MARK")
head(bun.I2M)
# convert converted 'individual' back to 'frequency' format
bun.I2F <- capHistConvert(bun.F2I,id="id",in.type="individual",
           out.type="frequency",var.lbls.pre="Sample")
head(bun.I2F)


########################################################################
## A larger example of 'marked' or 'RMark' format, but with a covariate
##   and when the covariate is removed there is no frequency or individual
##   fish identifier.
data(dipper,package="marked")
head(dipper)
# isolate males and females
dipperF <- subset(dipper,sex=="Female")
dipperM <- subset(dipper,sex=="Male")
# convert females to 'individual' format
dipF.R2I <- capHistConvert(dipperF,cols2ignore="sex",in.type="RMark")
head(dipF.R2I)
# convert males to 'individual' format
dipM.R2I <- capHistConvert(dipperM,cols2ignore="sex",in.type="RMark")
head(dipM.R2I)
# add sex variable to each data.frame and then combine
dipF.R2I$sex <- "Female"
dipM.R2I$sex <- "Male"
dip.R2I <- rbind(dipF.R2I,dipM.R2I)
head(dip.R2I)
tail(dip.R2I)

} # end interactive


## An example of problem with unused levels
## Create a set of test data with several groups
( df <- data.frame(fish=c("id17","id18","id21","id17","id21","id18","id19","id20","id17"),
                   group=c("B1","B1","B1","B2","B2","B3","B4","C1","C1")) )
#  Let's assume the user wants to subset the data from the "B" group
( df1 <- subset(df,group %in% c("B1","B2","B3","B4")) )
#  Looks like capHistConvert() is still using the unused factor
#  level from group C
capHistConvert(df1,id="fish",in.type="event")
# use droplevels() to remove the unused groups and no problem
df1 <- droplevels(df1)
capHistConvert(df1,id="fish",in.type="event")

Run the code above in your browser using DataLab