Reading: Final Project Overview

Estimated time: 4 minutes

Welcome to the final project of this course. This project is your opportunity to apply statistical analysis and predictive modeling techniques to a realistic healthcare operations scenario using Python and Jupyter Notebook.

You will work end-to-end with emergency department (ED) visit data—exploring patterns, testing hypotheses, and building models that support data-driven operational and clinical decision-making.

All analysis will be completed directly in this notebook by writing code in the designated cells.

Project scenario

You are a Healthcare Data Analyst working with the Emergency Department (ED) operations team at a large urban hospital.

ED leadership has raised concerns about:

• Prolonged length of stay (LOS) in the emergency department
• Differences in patient outcomes across triage severity levels
• The ability to identify patients likely to be admitted as early as possible

To support these goals, you have been asked to analyze historical ED visit data and build both statistical tests and predictive models that can inform operational planning and early risk stratification.

⚠️ Note: All data used in this project is synthetic and created solely for educational purposes.

Project objectives

By completing this project, you will demonstrate your ability to:

• Explore and summarize healthcare datasets using descriptive statistics
• Perform hypothesis testing to compare patient groups
• Use ANOVA to analyze differences across multiple triage categories
• Build and evaluate a logistic regression model for admission prediction
• Train and compare tree-based machine learning models
• Interpret analytical results in an operational healthcare context

About the dataset

You will work with a synthetic emergency department dataset containing 400 patient visits, with each row representing a single ED encounter. The data includes patient age, triage level, arrival time, length of stay, diagnostic tests ordered, and whether the patient was admitted.

The dataset will be loaded directly within this notebook, and no external data downloads are required.

What you'll do in this project

This project is divided into a series of deliverables, each representing a specific task. You will complete them sequentially in the given notebook.

You will:

1. Inspect the dataset structure and variable types
2. Summarize the emergency department length of stay
3. Visualize LOS differences by admission status
4. Perform hypothesis testing on LOS outcomes
5. Conduct ANOVA across triage severity levels
6. Build and evaluate a logistic regression model for admission risk
7. Evaluate model performance using ROC–AUC
8. Train and evaluate a decision tree classifier
9. Train and compare a random forest classifier
10. Interpret results to support ED operational decisions

Each deliverable includes a clearly marked section where you are expected to write and execute your own code.

How to complete and submit the project

In this final project, you will write and run Python code in the provided Jupyter Notebook (ED_Analytics_Project.ipynb) to complete a series of analytics and modeling tasks. You will use the notebook to develop, test, and validate your code, but you will not submit the notebook itself.

Instead, you will document your work in a provided project submission template (Project Submission Template_ED Data Analytics), where you will include:

• Written answers (where applicable)
• Code snippets or screenshots of your code
• Screenshots or outputs of results, tables, and plots

You can run and verify your code locally or using Google Colab to ensure all outputs are correct.

To submit your final project:

1. Complete all tasks in the notebook
2. Save the notebook file on your local machine
3. Rename the file using the following format:

ED Data Analytics_<YourInitials>.docx

Example:

ED Data Analytics_JD.docx

Estimated time

You should expect to spend approximately 1 hour and 15 minutes (project + submission document) completing this project and creating the submission document.

How will you be graded

While submitting, you can choose to have your submission evaluated by an AI tool or by a peer.

For AI-grading, you will receive your grades immediately after submission.

If you choose peer grading, you will receive your score after another learner evaluates your submission, and you complete the review of one peer's work using a simple rubric. You can also join an optional peer discussion to reflect and share takeaways.

Ready to begin?

This project mirrors the type of analysis performed by healthcare analytics teams working on real operational challenges.

When you're ready, proceed to the next section for detailed instructions and begin your analysis.

Good luck—and enjoy putting your skills into practice.