This project is a complete data engineering workflow that transforms the raw Titanic dataset into a clean, structured format optimized for Machine Learning models.

The goal was to move from "raw and messy" data to "AI-ready" data. This involves handling missing values, encoding categorical variables, and engineering new features to improve potential model accuracy.

• VS Code: Primary development environment.
• Python 3.10+: Core programming language.
• Pandas & NumPy: For data manipulation and numerical logic.
• Matplotlib & Seaborn: For exploratory data analysis (EDA).

• Age: Filled missing values using the median grouped by Passenger Class and Sex.
• Embarked: Filled missing values with the mode (most common port).
• Cabin: Dropped due to having over 77% missing data.

I created several new features to capture hidden patterns:

• Title: Extracted from names (Mr, Mrs, Miss, etc.).
• FamilySize: Combined SibSp and Parch.
• IsAlone: A binary flag for passengers traveling without family.
• FareBin & AgeBin: Grouped continuous data into logical categories.

• Converted Sex to binary (0/1).
• Applied One-Hot Encoding to Embarked and Title to ensure the data is 100% numerical.

• Dropped non-predictive columns: PassengerId, Name, Ticket, SibSp, and Parch.

• titanic_data_preparation.ipynb: The complete Jupyter Notebook with step-by-step logic.
• train.csv: The original raw dataset.
• titanic_clean_ai_ready.csv: The final processed output.

Completed as part of the AI Data Preparation Challenge.