In awatt008/R_Package_Wattler: Analysis of Survey Data

knitr::opts_chunk$set(echo = TRUE)

Installation

First,we're going to load in the libraries we will use with my functions:

install.packages("tidyverse")
install.packages("ggplot2")
library(tidyverse)
library(ggplot2)

Let's get my R package from github:

devtools::install_github("awatt008/R_Package_Wattler")
library(projects)

Some may need the surveys.csv data, so use this download/install code to get that:

download.file(url = "https://raw.githubusercontent.com/awatt008/R_Package_Wattler/master/data/surveys.csv", destfile = "/cloud/project/data/surveys.csv")

Now, we read the surveys data with tidyverse

surveys <- read_csv("/cloud/project/data/surveys.csv")

Functions

Let's take a look at my functions first, then we'll see how they're used with tidyverse:

 boxplot_weight_species <- function(., species_id,weight){
      if(!is.data.frame(surveys)){
        return("This is not a dataframe")
      }
     plt <- ggplot(., mapping = aes(x = weight, y = species_id)) +
            geom_boxplot()
            return(plt)
 }

The output of the boxplot_weight_species function plus the piped code will return a boxplot with weight and species selected and its corresponding data grouped from the surveys data frame.

plotted_weight_species<-surveys%>%
   filter(!is.na(species_id))%>%
   filter(!is.na(weight))%>%
   select(species_id, weight) %>% 
   boxplot_weight_species(species_id = species_id, weight = weight)

Next, we'll look at a function that removes NA's from surveys that only requires inputting the database name plus the function and weight column.

no_na_weight <- function(., weight){
  if(any(!is.na(surveys$weight)))
  scrub_weight_na <- surveys%>%
  filter(!is.na(weight))
return(scrub_weight_na)
}

So, this short line of code below will return surveys data without the NAs:

surveys%>%
no_na_weight(weight = weight)

And now, a function that creates a line of regression for surveys that works once we run a pipe that filters NA's, selects hindfoot and weight, and groups their corresponding data:

hindfoot_weight_regression <- function(., weight, hindfoot_length){
  if(!is.data.frame(surveys)){
    return("This is not a dataframe")
  }
  plt_regression <- ggplot(., aes(x = weight, y = hindfoot_length)) +
    geom_point(size = .25) + 
    geom_smooth(method = "lm",     
                color = "blue",    
                size = .5,        
                fill = "black")        
  return(plt_regression)
}

And here's that code:

plot_of_regression<- surveys%>%
filter(!is.na(weight))%>%
filter(!is.na(hindfoot_length))%>%
select(weight,hindfoot_length)%>%                 
group_by(weight,hindfoot_length)%>%
hindfoot_weight_regression(weight = weight, hindfoot_length = hindfoot_length)