Understanding Animal Adoption Through Data Science

6.5 million. That's how many animals end up in shelters each year. Out of that number, only about half are adopted including Sunshine, a domestic shorthair cat who found a loving home in Washington DC with Thinkful student Joanne Lin.

Every few weeks, Joanne who is currently enrolled in Thinkful's data science bootcamp, spends some free time browsing the internet for pictures of adoptable pets in her area. While she isn't quite ready to rescue another animal (she's a bit unsure if Sunshine would approve), she's been able to steer her passion for animals towards something of great importance: her Thinkful capstone project.

Is there a way to predict whether a pet gets adopted? This is the research question that Joanne posed before conducting her analysis. She used Kaggle to find her dataset (courtesy of the Austin Animal Center) and used the random forest classifier, a supervised learning algorithm to determine which features were the best predictors of adoptions.

According to Joanne's analysis, whether or not an animal fixed, the color and pattern of the animal's coat, and the age of the animal are among the top 5 predictive features for both cat and dog adoptions. However, whether or not the animal had a name was much more important for cat adoptions than dog adoptions. Conversely, the breed was much more important for dog adoptions than cat adoptions.

To read the rest of Joanne's findings including which dogs and cats are most "adoptable", check out the Thinkful blog.

This piece was sponsored by Thinkful

Want to learn more about Thinkful? Check out their alumni reviews on SwitchUp or read this great piece on Flexible Data Science.