🐍 Python Data Science Projects

This repository contains beginner-friendly Python data science projects, each with a Jupyter Notebook and a dedicated README**.

---

📁 Repository Structure

└── README.md                                    # ← You are here
└── Statistics.ipynb                             # Notebook: descriptive statistics & visualisation
└── Regression Analysis.ipynb                    # Notebook: head and brain prediction with ML
└── Predictive Analysis.ipynb                    # Notebook: house price prediction with ML
└── Logistic Regression SUV Predictions.ipynb    # Notebook: SUV purchase classification with Logistic Regression
└── Titanic Data Analysis.ipynb                  # Notebook: Titanic survival classification with Logistic Regression
├── Random_Forest_Classification_Penguins.ipynb  # Notebook: penguin species classification with Random Forest

---

📂 Projects at a Glance

1. 📊 Statistics Using Python

File: Statistics.ipynb

An introduction to descriptive statistics using a custom dataset and the classic Iris dataset.

What you'll learn	Tools used
Mean, Median, Mode	statistics (built-in)
Variance & Standard Deviation	statistics.pvariance, stdev
Summary statistics on real data	pandas .describe()
Pairwise feature visualisation	seaborn pairplot

💡 Key insight: Demonstrates why the median is more reliable than the mean when outliers are present — a right-skewed dataset makes this tangible.

---

2. 🧠 Linear Regression — Brain Weight Prediction

File: Linear Regression.ipynb

A hands-on implementation of Linear Regression from scratch, then validated using Scikit-learn — predicting brain weight from head size measurements.

What you'll learn	Tools used
Implementing regression manually (least squares)	numpy
Visualising regression lines & scatter plots	matplotlib
Calculating R² (coefficient of determination)	pandas
Validating with a ML library	sklearn.linear_model

💡 Key result: Built Linear Regression from scratch using the least squares formula, then confirmed the result with Scikit-learn — both approaches yielding the same R² score.

---

3. 🏠 Predictive Analysis — House Price Prediction

File: Predictive Analysis.ipynb

A complete machine learning pipeline that predicts house prices using Linear Regression, achieving a 92% R² score on test data.

What you'll learn	Tools used
Exploratory data analysis (EDA)	pandas
Visualising feature relationships	seaborn relplot
Train/test splitting	sklearn.model_selection
Building & evaluating a model	sklearn.linear_model

💡 Key result: The model explains 92% of variance in house prices — a strong fit demonstrating the power of Linear Regression on structured data.

---

4. 🚗 Classification — SUV Purchase Prediction

File: Logistic Regression SUV Predictions.ipynb

A complete classification pipeline that predicts whether a user will purchase an SUV based on age and salary, using Logistic Regression and achieving 89% accuracy on test data.

What you'll learn	Tools used
Feature selection by index	pandas / numpy
Train/test splitting	sklearn.model_selection
Feature scaling	sklearn.preprocessing
Building & evaluating a classifier	sklearn.linear_model

💡 Key result: The model correctly classifies 89% of unseen users as buyers or non-buyers — demonstrating Logistic Regression as a practical tool for binary classification problems.

---

5. 🚢 Classification — Titanic Survival Prediction

File: Titanic Data Analysis.ipynb

A complete data analysis and classification pipeline that explores the Titanic dataset and predicts passenger survival based on features like age, sex, and class, using Logistic Regression and achieving 79% accuracy on test data.

What you'll learn	Tools used
Exploratory data analysis (EDA)	pandas / numpy
Visualising survival patterns	seaborn countplot
Data cleaning & feature engineering	pandas
Building & evaluating a classifier	sklearn.linear_model
Confusion matrix & classification report	sklearn.metrics

💡 Key result: The model achieves 79% accuracy in predicting survival — uncovering how factors like passenger class, sex, and age influenced the odds of surviving the Titanic disaster.

---

6. 🐧 Classification — Penguin Species Prediction

File: Random Forest Classification_Penguins.ipynb

A complete multi-class classification pipeline that predicts penguin species (Adelie, Chinstrap, Gentoo) from physical measurements using Random Forest, achieving 98% accuracy on test data.

What you'll learn	Tools used
Data cleaning & null value handling	pandas
One-hot encoding of categorical features	pandas get_dummies
Train/test splitting	sklearn.model_selection
Building a Random Forest classifier	sklearn.ensemble
Tuning number of trees (n_estimators)	RandomForestClassifier

💡 Key result: With just 7 decision trees and Gini impurity as the criterion, the model reaches 98% accuracy — demonstrating how ensemble methods outperform single classifiers on structured biological data.

🙌 Acknowledgements

Built with Python and the open-source data science ecosystem: pandas · NumPy · seaborn · scikit-learn · matplotlib