Animal shelters provide a great service to the world by keeping animals safe and finding them a home that we provide them with love and affection. However the system is not perfect and sometimes the outcomes are less than desirable. Some shelters even resort to killing animals to make space for other animals. According to Peta, animal shelters face many issues of overcrowding, constant introduction of new, unknown animals into the “herd”, mandatory holding periods, and limited resources.
https://www.aspcapro.org/characteristics-challenges-shelter-environment https://www.peta.org/issues/animal-companion-issues/animal-shelters-hope-homeless/#:~:text=Common%20problems%20include%20cruel%20killing,inadequate%20screening%20procedures%20for%20adoption
I have collected data from the largest no kill animal shelter in the USA, Austin Animal Center. They provide care to 18,000 animals. With their data, we can study the relationship between their intake data compared to the outcome. Specifcally how does animal type, breed, gender, and age effect its outcome result. The goal of this project is to see how we can improve the outcome of animals being adopted.
If successful, this work may help animals have a better chance at a good outcome. We will analyze the animals that generally do not receive great outcomes and we can use it in our favor by marketing to the public about this issue and appealing to people's emotions.
One negative outcome of such a classifier is that it may encourage the shelters to only shelter dogs of certain breeds, age, and gender.
We will use a Kaggle Dataset of Austin Animal Center Shelter Intakes and Outcomes to observe 13 data columns regarding animal characteristics and adoption outcome:
This will be enough data to analyze how characterics of an animal will affect its time in the shelter and the outcome of their adoption.
import pandas as pd
data = pd.read_csv('aac_intakes_outcomes.csv')
data.head(10)
A potential problem this project may face is that these are only assumptions that these characterics may affect the outcome of their adoption. The data is also for animals strictly located in Austin, it may not be the same in other cities.
We will cluster the animals into subsets of animals with characteristics that describe their species, breed, age, gender, and color. Doing so will allow us to see which characteristics have a greater effect on the outcome of the adoption.