The problem statement addressed in this analysis pertains to understanding the factors influencing dog adoption rates. This is of interest as it can provide insights into improving the adoption process, enhancing animal welfare, and ultimately matching more dogs with suitable homes.
To address this problem statement, we plan to analyze the dog data set, which includes information on dog breeds, locations, adoption rates, and other relevant variables. We will employ descriptive statistics, exploratory data analysis, and potentially predictive modeling techniques to gain insights into the factors influencing dog adoption rates.
Our current proposed approach involves conducting exploratory data analysis to identify patterns and trends in the data. We will use techniques such as data visualization, correlation analysis, and possibly regression modeling to understand the relationships between various variables and adoption rates.
The analysis aims to provide valuable insights for those involved in dog adoption processes, including animal shelters, rescue organizations, and potential adopters. By understanding the factors influencing adoption rates, shelters and organizations can optimize their strategies for promoting dog adoptions, while potential adopters can make more informed decisions when choosing a dog. Ultimately, the analysis seeks to facilitate better matches between dogs and potential owners, leading to improved animal welfare outcomes.
Packages used in this analysis:
To suppress messages and warnings resulting from loading the package,
we can use the suppressMessages() function.
tidyr: This package provides tools for creating tidy data, where each column is a variable, and each row is an observation.
tidyverse: This is a collection of packages designed for data science, including dplyr, ggplot2, tidyr, and other packages. It provides a consistent and coherent approach to data manipulation, visualization, and analysis, making it easier to work with data.
dplyr: This function provides a set of functions that are optimized for tasks such as filtering, selecting, mutating, summarizing, and arranging data. It makes it easier to work with data frames and perform common data manipulation tasks efficiently.
Original source of data was obtained through Canvas Dog Data.
In the source data for the dog data set, there were originally 35 variables collected for the purpose of gathering information on animals who are up for adoption. The original purpose of the data set was likely to facilitate the adoption process by providing detailed information about the dogs available for adoption.
Regarding missing values, the data collection process encountered cases where certain information was not available or not recorded. In the dog data set, missing values were recorded as NA. As part of the data cleaning process, missing values were handled by encoding missing values as “Unknown”, and removing columns which had a high percentage of NA values that did not impact data calculations. In addition, variables that were not deemed useful to keep in the dataset were removed.
#load package
suppressPackageStartupMessages({
library(tidyverse)
library(dplyr)
})
# imported the data set
dog <- read.csv("dog_descriptions.csv")
# view structure of data
str(dog)
#summary of data
summary(dog)
# address missing data
#checked how many NA values each variable had
colSums(is.na(dog))
#removed 7 columns due to some having NA values that were 80% or more, and other columns had data that was unecessary
#Changed NA values to "unknown" in 8 columns to keep the columns and get rid of NA values
dog <- dog[, !(names(dog) %in% c("photo", "color_tertiary", "tags", "description","color_secondary", "accessed", "stateQ"))] %>%
mutate(env_children = ifelse(is.na(env_children), "Unknown", env_children)) %>%
mutate(env_dogs = ifelse(is.na(env_dogs), "Unknown", env_dogs)) %>%
mutate(env_cats = ifelse(is.na(env_cats), "Unknown", env_cats)) %>%
mutate(breed_secondary = ifelse(is.na(breed_secondary), "Unknown", breed_secondary)) %>%
mutate(color_primary = ifelse(is.na(color_primary), "Unknown", color_primary)) %>%
mutate(coat = ifelse(is.na(coat), "Unknown", coat)) %>%
mutate(contact_zip = ifelse(is.na(contact_zip), "Unknown", contact_zip)) %>%
mutate(type = ifelse(is.na(type), "Unknown", type))
# checking for data duplication
unique(dog)
# changed to only show the date in the posted column
dog$posted <- as.Date(dog$posted)Once the data had been cleaned, the following is what the final data set looks like in a condensed from showing the first 5 rows of the dataset:
head(dog, 5)## id org_id species breed_primary breed_secondary
## 1 46042150 NV163 Dog American Staffordshire Terrier Mixed Breed
## 2 46042002 NV163 Dog Pit Bull Terrier Mixed Breed
## 3 46040898 NV99 Dog Shepherd Unknown
## 4 46039877 NV202 Dog German Shepherd Dog Unknown
## 5 46039306 NV184 Dog Dachshund Unknown
## breed_mixed breed_unknown color_primary age sex size coat
## 1 TRUE FALSE White / Cream Senior Male Medium Short
## 2 TRUE FALSE Brown / Chocolate Adult Male Large Short
## 3 FALSE FALSE Brindle Adult Male Large Short
## 4 FALSE FALSE Unknown Baby Female Large Unknown
## 5 FALSE FALSE Unknown Young Male Small Long
## fixed house_trained declawed special_needs shots_current env_children
## 1 TRUE TRUE NA FALSE TRUE Unknown
## 2 TRUE TRUE NA FALSE TRUE Unknown
## 3 TRUE FALSE NA FALSE TRUE Unknown
## 4 FALSE FALSE NA FALSE FALSE Unknown
## 5 TRUE FALSE NA FALSE TRUE TRUE
## env_dogs env_cats name status posted contact_city contact_state
## 1 Unknown Unknown HARLEY adoptable 2019-09-20 Las Vegas NV
## 2 Unknown Unknown BIGGIE adoptable 2019-09-20 Las Vegas NV
## 3 Unknown Unknown Ziggy adoptable 2019-09-20 Mesquite NV
## 4 Unknown Unknown Gypsy adoptable 2019-09-20 Pahrump NV
## 5 TRUE TRUE Theo adoptable 2019-09-20 Henderson NV
## contact_zip contact_country type
## 1 89147 US Dog
## 2 89147 US Dog
## 3 89027 US Dog
## 4 89048 US Dog
## 5 89052 US DogThe following variables are variables of concern in the cleaned data set due to the number of “unknown” values.
##
## FALSE TRUE Unknown <NA>
## 4439 23588 30153 0##
## FALSE TRUE Unknown <NA>
## 3547 31122 23511 0##
## FALSE TRUE Unknown <NA>
## 6810 12542 38828 0To discover hidden insights in the data, I’ll start with Exploratory Data Analysis. This means diving into the data to understand its patterns, distributions, and connections. I’ll use different graphs like histograms, box plots, scatter plots, and correlation matrices to spot any unusual patterns or outliers. To answer key questions, I would summarize the data using descriptive statistics, frequency tables, and summary metrics. Additionally, I could use inferential statistics such as hypothesis testing or regression analysis to test relationships and make predictions. Overall, by leveraging a combination of exploratory techniques, data manipulation, and statistical analysis, I aim to uncover hidden insights in the data and provide valuable information that can inform decision-making processes related to dog adoption and welfare.
To explain what I find in the data, I’ll use different pictures and tables. For example:
In addition, some tables that may be used are:
Some things I currently don’t know or would need extra assistance in to answer my questions is conducting exploratory analysis with the current variables and conducting descriptive statistics. I am not quite comfortable in using these techniques, especially using the RStudio software. I am unfamiliar in conducting statistical work in the software and feel like it may hinder me from being able to conduct helpful and meaningful test on the dataset.
I do not plan to incorporate any machine learning techniques as I am not familiar or have knowledge on how to apply those techniques.