Colorado Parks and Wildlife (CPW) aka Colorado Department of Wildlife (CDOW) provides historical statistics on Big Game hunts. The tables are organized by year and available for download in pdf format. In the case of the project we are solely interested in Elk Rifle Hunting on public land. The tables are organized by hunting seasons (First-Fourth), as well as by hunting regions (Units) The hunting regions vary very slightly from year to year but for the most part have been consistent for many years.
setwd("~/_code/colorado-dow/datasets")Load required libraries for acquiring data from pdf
library(pdftools,quietly = T)
library(stringr,quietly = T)
# Identify the years that CDOW will provide tables for in this pdf format
years <- c(2006,2007,2008,2009,2010,2011,2012,2013,2014,2015,2016,2017)Loop through years
COElkRifleAll <- NULL # Initialize
for (iyear in years) {
# RUN ONCE to download
# if (iyear >= 2014) {
# download.file(paste("http://cpw.state.co.us/Documents/Hunting/BigGame/Statistics/Elk/",
# iyear,"StatewideElkHarvest.pdf",sep=""),
# paste(iyear,"COElkHarvest",sep=""))
# } else {
# download.file(paste("http://cpw.state.co.us/Documents/Hunting/BigGame/Statistics/Elk/",
# iyear,"ElkHarvestSurvey.pdf",sep=""),
# paste(iyear,"COElkHarvest",sep=""))
# }
# This function will directly export the raw text in a character vector with spaces to show
# the white space and \n to show the line breaks.
COElk <- pdf_text(paste(iyear,"COElkHarvest",sep=""))
# Having a full page in one element of a vector is not the most practical. Using strsplit
# will help separate lines from each other
COElka <- strsplit(COElk, "\n")
# years starting in 2014 have a cover page or table of contents
if (iyear >= 2014) {
COElka <- COElka[-1] # remove cover page (map, or table of contents)
}
# The document holds more information than we are after.
# In our case we are looking for
tableheadings <- c(paste(iyear,"Elk Harvest, Hunters and Recreation Days for First Rifle Seasons"),
paste(iyear,"Elk Harvest, Hunters and Recreation Days for Second Rifle Seasons"),
paste(iyear,"Elk Harvest, Hunters and Recreation Days for Third Rifle Seasons"),
paste(iyear,"Elk Harvest, Hunters and Recreation Days for Fourth Rifle Seasons"))
# Notice that the tables run across pages, there are some pages that will have
# info we want and info we want to ignore
# Identify pages with the table headings we identified
rifle1 <- grep(tableheadings[1], COElka)
rifle2 <- grep(tableheadings[2], COElka)
rifle3 <- grep(tableheadings[3], COElka)
rifle4 <- grep(tableheadings[4], COElka)
rifleseasons <- unique(c(rifle1,rifle2,rifle3,rifle4))
COElkb <- COElka[rifleseasons]
# the first page has the end of a previous table, remove it so we can have consistent columns
firsttable <- COElkb[[1]]
# which row has the heading of our table?
firsttablestart <- grep(tableheadings[1], firsttable)
# drop all rows before that table
firstpage <- firsttable[-(1:firsttablestart-1)]
# the last page might have the beginning of a new table, remove it so we can have consistent columns
lasttablepage <- length(COElkb) # what is the last page?
lasttable <- COElkb[[lasttablepage]]
# which rows have table headings? they all start with 'iyear Elk Harvest'
lasttableend <- grep(paste(iyear,"Elk Harvest"), lasttable)
lasttableend <- max(lasttableend) # the second entry is the start of the table to ignore
if (lasttableend >1) {
# drop all rows after our table
lastpage <- lasttable[-(lasttableend:length(lasttable))]
} else {lastpage <- lasttable}
# replace updated first and last pages
COElkc <- COElkb
COElkc[[1]] <- firstpage
COElkc[[lasttablepage]] <- lastpage
# unlist page elements
COElkd <- unlist(COElkc)
######## SEASON ONE ######## (could make a season loop as well)
# identify season 1 data
seasonONEstart <- grep(tableheadings[1], COElkd)[1]
seasonONEend <- grep(tableheadings[2], COElkd)[1]
seasonONE <- COElkd[((seasonONEstart+1):(seasonONEend-1))]
# remove rows with the table headers
removeheaderrows <- grep(paste(iyear,"Elk Harvest"), seasonONE)
seasonONE <- seasonONE[-removeheaderrows]
# determine column names
columnnames <- grep("([:alpha:])", seasonONE)
seasonONE1 <- seasonONE[-columnnames]
columnnames1 <- seasonONE[columnnames]
columnnames1 <- columnnames1[-length(columnnames1)] # we know the last row is a 'Total' summary, not a name
columnnames2 <- columnnames1[length(columnnames1)]
columnnames2 <- str_trim(columnnames2) # remove any extra whitespace
columnnames3 <- unlist(strsplit(columnnames2,split = "\\s+"))
colremove <- grep("([.])",columnnames3)
columnnames3 <- columnnames3[-colremove]
seasonONE1 <- str_trim(seasonONE1) # remove extra whitespace
pagenumbers <- grep("\\s+", seasonONE1) # remove page numbers
seasonONE2 <- seasonONE1[pagenumbers]
# now that it is cleaned up, use the white space to separate into columns
seasonONE3 <- str_split_fixed(seasonONE2,pattern = "\\s+", n=length(columnnames3))
seasonONE3 <- as.data.frame(seasonONE3) # and convert into a dataframe
colnames(seasonONE3) <- columnnames3 # apply our column names
# add the season column
seasonONE3$Season <- 1
######## SEASON TWO ########
# identify season 2 data
seasonTWOstart <- grep(tableheadings[2], COElkd)[1]
seasonTWOend <- grep(tableheadings[3], COElkd)[1]
seasonTWO <- COElkd[((seasonTWOstart+1):(seasonTWOend-1))]
# remove rows with the table headers
removeheaderrows <- grep(" Elk Harvest", seasonTWO)
seasonTWO <- seasonTWO[-removeheaderrows]
# determine column names
columnnames <- grep("([:alpha:])", seasonTWO)
seasonTWO1 <- seasonTWO[-columnnames]
columnnames1 <- seasonTWO[columnnames]
columnnames1 <- columnnames1[-length(columnnames1)] # we know the last row is a 'Total' summary, not a name
columnnames2 <- columnnames1[length(columnnames1)]
columnnames2 <- str_trim(columnnames2) # remove any extra whitespace
columnnames3 <- unlist(strsplit(columnnames2,split = "\\s+"))
colremove <- grep("([.])",columnnames3)
columnnames3 <- columnnames3[-colremove]
seasonTWO1 <- str_trim(seasonTWO1) # remove extra whitespace
pagenumbers <- grep("\\s+", seasonTWO1) # remove page numbers
seasonTWO2 <- seasonTWO1[pagenumbers]
# now that it is cleaned up, use the white space to separate into columns
seasonTWO3 <- str_split_fixed(seasonTWO2,pattern = "\\s+", n=length(columnnames3))
seasonTWO3 <- as.data.frame(seasonTWO3) # and convert into a dataframe
colnames(seasonTWO3) <- columnnames3 # apply our column names
# add the season column
seasonTWO3$Season <- 2
######## SEASON THREE ########
# identify season 3 data
seasonTHREEstart <- grep(tableheadings[3], COElkd)[1]
seasonTHREEend <- grep(tableheadings[4], COElkd)[1]
seasonTHREE <- COElkd[((seasonTHREEstart+1):(seasonTHREEend-1))]
# remove rows with the table headers
removeheaderrows <- grep(" Elk Harvest", seasonTHREE)
seasonTHREE <- seasonTHREE[-removeheaderrows]
# determine column names
columnnames <- grep("([:alpha:])", seasonTHREE)
seasonTHREE1 <- seasonTHREE[-columnnames]
columnnames1 <- seasonTHREE[columnnames]
columnnames1 <- columnnames1[-length(columnnames1)] # we know the last row is a 'Total' summary, not a name
columnnames2 <- columnnames1[length(columnnames1)]
columnnames2 <- str_trim(columnnames2) # remove any extra whitespace
columnnames3 <- unlist(strsplit(columnnames2,split = "\\s+"))
colremove <- grep("([.])",columnnames3)
columnnames3 <- columnnames3[-colremove]
seasonTHREE1 <- str_trim(seasonTHREE1) # remove extra whitespace
pagenumbers <- grep("\\s+", seasonTHREE1) # remove page numbers
seasonTHREE2 <- seasonTHREE1[pagenumbers]
# now that it is cleaned up, use the white space to separate into columns
seasonTHREE3 <- str_split_fixed(seasonTHREE2,pattern = "\\s+", n=length(columnnames3))
seasonTHREE3 <- as.data.frame(seasonTHREE3) # and convert into a dataframe
colnames(seasonTHREE3) <- columnnames3 # apply our column names
# add the season column
seasonTHREE3$Season <- 3
######## SEASON FOUR ########
# identify season 4 data
seasonFOURstart <- grep(tableheadings[4], COElkd)[1]
seasonFOUR <- COElkd[((seasonFOURstart+1):(length(COElkd)))] # to the end
# remove rows with the table headers
removeheaderrows <- grep(" Elk Harvest", seasonFOUR)
seasonFOUR <- seasonFOUR[-removeheaderrows]
# determine column names
columnnames <- grep("([:alpha:])", seasonFOUR)
seasonFOUR1 <- seasonFOUR[-columnnames]
columnnames1 <- seasonFOUR[columnnames]
columnnames1 <- columnnames1[-length(columnnames1)] # we know the last row is a 'Total' summary, not a name
columnnames2 <- columnnames1[length(columnnames1)]
columnnames2 <- str_trim(columnnames2) # remove any extra whitespace
columnnames3 <- unlist(strsplit(columnnames2,split = "\\s+"))
colremove <- grep("([.])",columnnames3)
columnnames3 <- columnnames3[-colremove]
seasonFOUR1 <- str_trim(seasonFOUR1) # remove extra whitespace
pagenumbers <- grep("\\s+", seasonFOUR1) # remove page numbers
seasonFOUR2 <- seasonFOUR1[pagenumbers]
# now that it is cleaned up, use the white space to separate into columns
seasonFOUR3 <- str_split_fixed(seasonFOUR2,pattern = "\\s+", n=length(columnnames3))
seasonFOUR3 <- as.data.frame(seasonFOUR3) # and convert into a dataframe
colnames(seasonFOUR3) <- columnnames3 # apply our column names
# add the season column
seasonFOUR3$Season <- 4
#Combine
COElkRifle <- rbind(seasonONE3,seasonTWO3,seasonTHREE3,seasonFOUR3)
COElkRifle$Year <- as.character(iyear)
COElkRifleAll <- rbind(COElkRifleAll,COElkRifle)
}Clean up dataframe fields It may happen that when reading numeric data into R (usually, when reading in a file), they come in as factors. If f is such a factor object, you can use
# as.numeric(as.character(f)) to get the numbers back. More efficient, but harder to remember, is
# as.numeric(levels(f))[as.integer(f)]However, there are still implications in our data as there are commas present for values in the thousands. As-is converting to numeric using the factor levels will coerce them to NAs So we will need to convert to character, remove the commas, then convert to numeric
COElkRifleAll$Hunters <- as.numeric(gsub(",", "", COElkRifleAll$Hunters))
COElkRifleAll$Days <- as.numeric(gsub(",", "", COElkRifleAll$Days))
# TODO make the gsub search each column
COElkRifleAll$Success <- as.numeric(levels(COElkRifleAll$Success))[as.integer(COElkRifleAll$Success)]
COElkRifleAll$Harvest <- as.numeric(levels(COElkRifleAll$Harvest))[as.integer(COElkRifleAll$Harvest)]
COElkRifleAll$Season <- as.character(COElkRifleAll$Season)
COElkRifleAll$Unit <- as.character(COElkRifleAll$Unit)Create new statistics based on investigation
# How much effort it takes to have a successful result
COElkRifleAll$Harvest_Effort <- COElkRifleAll$Days / COElkRifleAll$Harvest # From Phase I investigation
COElkRifleAll$Harvest_Effort[is.infinite(COElkRifleAll$Harvest_Effort)] <- NA # we get an inf where there was no harvestPeek at the dataframe
head(COElkRifleAll)## Unit Bulls Cows Calves Harvest Hunters Success Days Season Year
## 1 411 0 0 0 0 26 0 79 1 2006
## 2 42 34 3 0 37 125 30 353 1 2006
## 3 421 44 38 0 82 351 23 1236 1 2006
## 4 43 56 39 9 104 472 22 1699 1 2006
## 5 44 53 37 10 100 474 21 1485 1 2006
## 6 444 17 39 5 61 285 21 892 1 2006
## Harvest_Effort
## 1 NA
## 2 9.540541
## 3 15.073171
## 4 16.336538
## 5 14.850000
## 6 14.622951