Predicting the Outcome of Dogs at the Austin Animal Center

Presented by:
Giulia Bronzi, Hannah Kim, Willa Sun, Nikki Tong (Wellesley College)
Abstract:

The Austin Animal Center is the largest no-kill animal shelter in the United States. This paper uses intake and outcome data on animals entering the shelter, and looks at variables such as breed, size, and sex to predict the outcome of dogs. After dividing the data into training and testing subsets, we created a regression tree that, based on predictors such as age, size based on breed, sex, whether the animals were spayed or neutered, condition upon intake, and the time they spent in the shelter, predicts whether a dog will be adopted, returned, transferred, or has died. From the tree, we can see that age at outcome is the most important predictor, consistent with the intuition that younger animals tend to be adopted more easily. For dogs that are older upon outcome, intake condition has a high predictive power. When income conditions are normal, adoption preferences seem consistent. Dogs that were younger upon intake and smaller size have a higher likelihood of getting adopted. We hope that these results can be used to help inform shelters' decisions upon the intake of a new animal.