Medical Appointment No-Shows: Data Cleaning & Analysis

This project focuses on cleaning and preparing a real-world healthcare dataset that records patient appointment details and whether or not they showed up. The data is sourced from Kaggle's "Medical Appointment No Shows" dataset and includes patient demographics, health conditions, and scheduling information.

Dataset Overview

Source: Kaggle - Medical Appointment No Shows

The dataset includes 110,000+ rows with the following key columns:

• PatientId: Unique identifier for each patient
• Gender, Age, Neighbourhood
• Scholarship, Hypertension, Diabetes, Alcoholism, Handicap
• SMS_received: Whether the patient received a reminder
• ScheduledDay, AppointmentDay
• No-show: Whether the patient showed up or not

---

Task 1: Data Cleaning & Preparation

Cleaning Goals:

• Rename columns with typos and inconsistent casing (e.g., Hipertension → Hypertension)
• Convert date/time fields to datetime format
• Handle invalid age values (e.g., negative numbers)
• Standardize categorical formats (e.g., No-show → 0/1)
• Remove duplicates and nulls
• Feature Engineering:
  • Waiting time (WaitDays) between scheduling and appointment
  • Appointment day of the week