🐾 ML_Zoomcamp_2025 – Austin Animal Center Project

Problem Statement -What we’re trying to predict:
This project predicts whether a pet entering the Austin Animal Center will be adopted or not. Using information such as animal type, breed, intake condition, and other relevant attributes, the model outputs a probability of adoption for each animal.

1. Dataset- data folder

This project uses public datasets from the Austin Animal Center, which include:
• Intakes: Records of animals entering the shelter
Outcomes: Records of animals leaving the shelter

These datasets contain information such as animal type, breed, intake reason, outcome type, and dates.

2. ,EDA , Data Cleaning
   The notebook performs the following preprocessing steps:
   • ✅ Loads both datasets and merges them on animal_ID
   • ✅ Filters out irrelevant columns and handles missing values
   • ✅ Converts date columns to datetime format
   • ✅ Encodes categorical features (e.g., animal type, intake condition)
   • ✅ Creates a binary target column: (e.g., Adopted vs. Not Adopted)
   

Feature Selection : categorical = ["intake_type", "intake_condition", "animal_type", "sex_upon_intake", "breed", "color"]
numeric = ["month", "year"]

3. Model
   The model used is a Random Forest Classifier, trained to predict the outcome of an animal based on:
   • Animal type and breed
   • Intake condition and type
   • Age upon intake
   • Time spent in shelter
   Output: Predicted outcome category (e.g., Adopted, Returned to Owner)
   

4. Train the Model
   To train the model, run the notebook . It will:
   • Load and clean the data
   • Engineer features and encode categories
   • Split into training and test sets
   • Train the model and evaluate performance
   

5. Make Predictions
   
   The notebook includes examples of predicting outcomes for new animal entries. You can modify the input features to test different scenarios.
   

6. FastAPI Service
   
   To deploy the model as an API:
   

• Interactive docs: http://localhost:8000/docs

7. Docker Deployment
   To containerize the API:

docker build -t pet-adoption-fastapi .
docker run -p 8000:8000 pet-adoption-fastapi

API will be available at http://localhost:8000/docs
Health check: curl http://localhost:8000/health