Project Info

• Client: Data Science with IBM
• Languages: Python
• Platform: Jupyter Notebook
• Completed: Sep 2021

This project addressed the critical issue of patient no-shows in healthcare, a challenge that disrupts resource allocation and limits access to care. Using a dataset of 110,000 appointments from public healthcare institutions in a Brazilian city, data cleaning, feature engineering, and exploratory data analysis were applied to optimize the data. Various supervised learning algorithms were tested, evaluated by metrics such as Accuracy, Precision, Recall, and F1 Score. The Random Forest classifier emerged as the top performer, achieving a 70% accuracy score. These results provide a foundation for healthcare providers to develop effective no-show prediction models, enhancing patient care and operational efficiency.