In brouwern/DRmencia: Data and code to replicate analyses for Latta, Brouwer, et al. 2018. PeerJ.

knitr::opts_chunk$set(echo = TRUE)

General Workflow

Data was recieved from Steve Latta as seperate spreadsheets which were merge into a single .csv file in 0_proofing_merging....Rmd script

This script tidies/cleans up the data.

The following script makes final changes to prepare for analysis

General Description of what this script does

See 0_proofing_merging....Rmd for more details

SCRIPT 1: reshaping, tidying, and cleaning data

The following steps aren't necessarily all distinct in practice or necessarily occur in this precise order

Step 1.1: "reshaping" and "tidying"

"reshape"

taking data from the format it was entered in into the format that works best w/ R; usually reshaping from "wide" to "long" format. eg, with data from each subplot in a different column to all data in a single column with subplot number indicated in a desperate column.

"tidying"

converting into an analysis-read format. For example, splitting things like columns that combine treatment and plot number into separate columns. (treatment column, subplot column)

Step 1.2: "cleaning"

Taking reshaped data and fixing any issues just as typos, incorrect column names, data entered into wrong field, fixing species names, etc.

After cleaning and tidying, the data could be "reshaped" (or recast) back into its original format and there would still be an ~1:1 correspondence with the datasheet.

Libraries

library(reshape2)
library(stringr)
library(stringi)
library(lubridate)
library(here)

Load data

Set up working directories

Load file

filename <- here::here("data-raw",
                  "data-raw-mencia",
                  "mencia_site_persistence",
                  "CSV_mencia_site_persistence",
                 "mencia_all_b.csv")
print(filename)

skip <- 7
dat <- read.csv(file = filename,
                skip = skip,
                header =TRUE,
                strip.white = TRUE,
                blank.lines.skip = TRUE,
                na.strings = c("NA",""," "),
                stringsAsFactors = FALSE,
                colClasses=c("SUF"="character")
                )

Check data

Check head and tail

Number blank columns to rigth and below spreadsheet

head(dat); tail(dat)

Check summary

summary(dat)

Check dim()

dim(dat)

Change all headings to lower case

names(dat) <- tolower(names(dat))

Reconcile names w/ other dataset

names(dat[1:10])
names(dat) <- gsub("pre","band.pre",names(dat))
names(dat) <- gsub("suf","band.suf",names(dat))

Set up year variable

summary(factor(dat$date))

dat$date2 <- dmy(dat$date)

dat$year <- year(dat$date2)
dat$month <- month(dat$date2)
dat$day <- day(dat$date2)

Move "died" from "wander" column to notes

dat$notes <- ifelse(dat$wanderer == "DIED","DIED",NA)

Wander column

summary(factor(dat$wanderer))
dat$wanderer <- gsub("X",NA,dat$wanderer)
dat$wanderer <- gsub("DIED",NA,dat$wanderer)
dat$wanderer <- factor(dat$wanderer)

summary(factor(dat$wanderer))

Add "La" to sites

dat$site <- gsub("Corral","El Corral",dat$site)
dat$site <- gsub("Caoba","La Caoba",dat$site)
dat$site <- gsub("Cueva","La Cueva",dat$site)

Convert to factors

dat$band.pre <- factor(dat$band.pre)
dat$band.suf <- factor(dat$band.suf)
dat$colors <- factor(dat$colors)
dat$age <- factor(dat$age)
dat$year.num <- dat$year
dat$year <- factor(dat$year)
dat$site <- factor(dat$site)

Check out data

summary(dat)

summary(factor(dat$sex),1000)

summary(factor(dat$notes),1000)

Sites

Fix typos

dat$colors <- factor(gsub("x-","X-",as.character(dat$colors)))

dat$species <- factor(gsub("AM-KE","AMKE",as.character(dat$species)))

dat$sex <- factor(ifelse(dat$sex == "U", NA,dat$sex))

Calculate age of sites

site.age.tab <- data.frame(site = levels(dat$site),
                      site.age.init = c(20,5,2,10))

dat <- merge(dat,site.age.tab)

dat$site.age <- dat$site.age.init + 
                      dat$year.num -2003

Create unique ID

dat$band <- with(dat, paste(band.pre,
                                  band.suf,
                                  sep = "-"))

Code spp

These are species judge to have sufficient sample sizes in 1st MS

Focal migrants

focal.mig <- c("OVEN","BAWW","COYE","AMRE","CMWA",
               "BTBW","PAWA","PRAW")

Focal res

focal.res <- c("HLCU"
,"STOF"
,"RLTH"
,"NOMO"
,"GRWA"
,"BANA"
,"BCPT"
,"YFGR"
,"BFGR"
,"GABU")

Code focal spp as resident or migrant

i1 <- which(dat$species %in% focal.mig)
i2 <- which(dat$species %in% focal.res)

dat$status.focals <- NA
dat[i1, "status.focals"] <- "mig"
dat[i2, "status.focals"] <- "res"

CHange "date" to "date.cap.orig"

names(dat)[grep("date2",names(dat))] <- "date.cap.orig"

Shorten names

names(dat) <- gsub("resight","rs",names(dat))
names(dat) <- gsub("recap","rc",names(dat))
names(dat) <- gsub("20","",names(dat))
names(dat) <- gsub("nov","N",names(dat))
names(dat) <- gsub("jan","J",names(dat))
names(dat) <- gsub("feb","F",names(dat))
names(dat) <- gsub("wanderer","wander",names(dat))
names(dat) <- gsub("wander","wndr",names(dat))
names(dat) <- gsub("species","spp",names(dat))
names(dat) <- gsub("wing","wng",names(dat))
names(dat) <- gsub("tarsus","tar",names(dat))
names(dat) <- gsub("year","yr",names(dat))
names(dat) <- gsub("month","mo",names(dat))
names(dat) <- gsub("band","bnd",names(dat))
names(dat) <- gsub("sex","sx",names(dat))
names(dat) <- gsub("colors","col",names(dat))
names(dat) <- gsub("notes","nts",names(dat))
names(dat) <- gsub("status","stat",names(dat))
names(dat) <- gsub("age","ag",names(dat))

Change "n/a" to "bf" for "before" 1st capture

Steve has "n/a" list for sampling occassions w/in a season before a bird was 1st observe (but not for years b/f)/

Change these to "bf"

natobf<- function(x){ifelse(x == "n/a","bf",x)}
j.r <- grep("^r", names(dat))

dat[,j.r] <- apply(dat[,j.r],2, natobf)

Change "X" to 1

steve list "X" for an observation or recapture.
change to 1 to be consistent the mark recaput

Xto1<- function(x){ifelse(x == "X" | x == "x","1",x)}
j.r <- grep("^r", names(dat))

dat[,j.r] <- apply(dat[,j.r],2, Xto1)

Change site names to 1

steve list site names in some cases instead of "X".

Change to 1

siteto1<- function(x){
  i <- which(x %in% c("Corral","Caoba","Cueva","Morelia","Cue/Cao",
                      "Fuera","DIED",
                      "repeated combo"))
  x[i] <- 1
  return(x)
  }

dat[,j.r] <- apply(dat[,j.r],2, siteto1)

Convert all to factor

for(i in 1:length(j.r)){
  j <- j.r[i]
  dat[,j] <- as.factor(dat[,j])
}
summary(dat[,j.r])
summary(factor(dat$rs.04.J))

Calculate number of times captured

calc_caps <- function(x){
  return(sum(as.numeric(x[grep("1",x)])))
}

dat$num.rc.rs <- apply(dat[,j.r],1, calc_caps)

i.drop1 <- which(names(dat) %in% c("bnd.pre",#"bnd.suf",
                                       "wng","tar",
                                       "site.ag.init",
                                       "site.ag","bnd",
                                       "nts","yr.num",
                                       "date2","mo","day"))

dat[,-c(j.r,i.drop1)]

Were any wanders incorrectly coded?

None

which(dat$wndr == "W" & dat$num.rc.rs > 0)

Were any P (=persistent) never resighted or recaptured?

nope

which(dat$wndr == "P" & dat$num.rc.rs < 1)

Are all residents coded

Yes

i.prob <- which(dat$wndr == "P" & dat$num.rc.rs < 1)

#dat[i.prob, -c(j.r,i.drop1)]

Examine wanderer

summary(factor(dat$wndr ))

i.w <- which(dat$wndr == "W")
j.r <- grep("^r", names(dat))

Code "P"

dat$persist <- ifelse(dat$num.rc.rs > 0, "P","NP")

Change spp to spp.code

names(dat) <- gsub("spp","spp.code",names(dat))

Load spp meta / trait info

spp.meta <- read.csv(here::here("data","spp_list.csv"),stringsAsFactors =F)

Subset columns I want to use

spp.meta <- spp.meta[c("spp.code","spp",
                       "status2",
                       "hab1","diet")]

names(spp.meta) <- gsub("status2","stat",names(spp.meta))

summary(spp.meta)

Change GRWA -> GTGT

Green tailed ground warlber and green tailed ground tanager are the same

Change column name

names(dat) <- gsub("species","spp.code", names(dat))

dat$spp.code <- as.character(dat$spp.code)
dat$spp.code[which(dat$spp.code == "GRWA")] <- "GTGT"

dat$spp.code[which(dat$spp.code == "HLCU")] <- "HILC"

MYWA = myrtyle warbler = YRWA yellow rumped warbler NUMA -> SBMU #nutmet manakin = scaly breasted munia WHQD = WFQD

dat$spp.code[which(dat$spp.code == "MYWA")] <- "YRWA"
dat$spp.code[which(dat$spp.code == "NUMA")] <- "SBMU"
dat$spp.code[which(dat$spp.code == "WHQD")] <- "WFQD"

Merge cleaned data w/ species meta data

Merge

dat2 <- merge(dat, spp.meta, all = T)

dim(dat)
dim(dat2)

dat2 <-dat2[-which(is.na(dat2$site) == TRUE), ]
dat <- dat2

Save cleaned data in original form

Save .csv

filename <- "mencia_site_persist_clean.csv"
filename <- here("data-raw",
                  "data-raw-mencia",
                  "mencia_site_persistence",
                  filename)

write.csv(dat, file = filename,row.names = F)

brouwern/DRmencia documentation built on May 6, 2019, 12:24 p.m.