Titanic EDA – Exploratory Data Analysis

Author: Tirtha Dutta
Date: 24 June 2025
Dataset: Kaggle – Yasser H Titanic Dataset

---

Objective

Explore the Titanic dataset using visual and statistical techniques to uncover relationships, trends, and anomalies.
This step helps lay the foundation for building accurate machine learning models.

---

Key EDA Steps Performed

1. Summary statistics for all numerical features
2. Histograms to visualize distributions
3. Boxplots to inspect outliers and spread
4. Correlation heatmap to check relationships
5. Markdown cell summarizing key findings

---

How to run locally?

git clone https://github.com/tirtha103/titanic-eda.git cd titanic-eda

Create and activate virtual environment (Windows)

python -m venv venv venv\Scripts\activate

Install dependencies

pip install -r requirements.txt

Launch the notebook

jupyter lab notebooks/01_eda_walkthrough.ipynb

---

Folder Structure

titanic-eda/
│
├── data/
│   └── titanic_cleaned.csv                # Cleaned dataset (from Task 1)
│
├── images/
│   ├── histograms.png                     # Histograms of numeric features
│   ├── boxplots.png                       # Boxplots for outlier inspection
│   └── correlation_heatmap.png            # Correlation heatmap
│
├── notebooks/
│   └── 01_eda_walkthrough.ipynb           # Full EDA notebook
│
├── report/
│   └── titanic_eda_report.pdf             # Optional PDF report
│
├── requirements.txt                       # Python package requirements
├── LICENSE                                # MIT License
└── .gitignore                             # Ignored files