Spaceship-Titanic-and-Decision-Trees

This project is a two-part exploration of the Spaceship Titanic dataset, a fictional dataset inspired by the historical Titanic disaster, available on Kaggle.

The documentation is a part of the jupyter notebook. A summary of the project is provided here. My blog post, with a concise & non-technical explanation can be found here.

---

Part 1: Exploratory Data Analysis (EDA)

• Utilizes the Pandas library for data manipulation and cleaning.
• Leverages Matplotlib and Seaborn for data visualization to uncover patterns and relationships within the dataset.
• This initial analysis provides insights into the characteristics of passengers and the overall journey that might influence survival outcomes.

---

Part 2: Implementing a Decision Tree Classifier from Scratch

• This section focuses on building the Decision Tree classification algorithm from scratch using only NumPy arrays.
• Our decision tree is trained on the Spaceship Titanic dataset to predict passenger survival based on relevant features.
• Our model is compared with sklearn's DecisionTreeRegressor and performs better by 0.05 accuracy points.