Deep Learning, Machine Learning & Data Analytics Sandbox

This repository is a structured Data Science portfolio covering the full analytical pipeline — from NumPy foundations and data cleaning through classical ML, unsupervised learning, and deep learning architectures.

---

📁 Repository Structure

ds-analytics-sandbox/
├── data/               ← Raw .csv datasets used across notebooks
├── notebooks/          ← All Jupyter notebooks (see modules below)
├── pyproject.toml
├── poetry.lock
└── README.md

---

🔹 Module 1: Foundations — NumPy & EDA

Core mathematical tools and exploratory data analysis.

File	Description
numpy_foundations.ipynb	NumPy basics — vectors, matrices, array operations
eda_bestsellers_analysis.ipynb	Exploratory data analysis on Amazon bestsellers dataset

---

🔹 Module 2: Data Engineering — Pandas & SQL

Data cleaning, transformation, and aggregation pipelines.

File	Description
pandas_data_cleaning.ipynb	Data ingestion, missing value imputation, type casting
pandas_dataframe_transformations.ipynb	Table reshaping, merging, joining strategies
pandas_metrics_aggregation.ipynb	Multi-source analytics, grouping, pivoting, statistical summaries

---

🔹 Module 3: Classical Machine Learning

Supervised and unsupervised learning pipelines.

File	Description
linear_regression_from_scratch.ipynb	Vectorized linear regression with gradient descent from scratch (NumPy)
ml_regularization_and_tuning.ipynb	Overfitting diagnostics, L1/L2 regularization, bias-variance tradeoff
ml_classification_models.ipynb	Supervised classifiers (activity recognition dataset), feature scaling, confusion matrices
ml_unsupervised_clustering.ipynb	KMeans clustering, Elbow method, market segmentation

---

🔹 Module 4: Deep Learning

Neural network architectures built with TensorFlow/Keras.

File	Description
mlp_low_level_tensorflow_mnist_colab.ipynb	Low-level TensorFlow MLP on MNIST (manual training loop, GradientTape)
dense_network_fashion_mnist_colab.ipynb	Fully-connected network on Fashion MNIST — hyperparameter experiments, Dropout, BatchNorm
cnn_vgg16_fashion_mnist_colab.ipynb	CNN + Transfer Learning (VGG16) on Fashion MNIST — feature extraction & fine-tuning
rnn_lstm_imdb_sentiment_colab.ipynb	RNN/LSTM/Bidirectional architectures on IMDB sentiment classification

---

🔹 Module 5: NLP & Recommendation Systems

Text processing and collaborative filtering.

File	Description
nlp_text_summarization_spacy.ipynb	Text summarization with spaCy and NLTK (extractive, frequency-based)
recommendation_systems_movielens.ipynb	Collaborative filtering with SVD, SVD++, NMF on MovieLens 100k

---

🛠️ Tech Stack

• Languages: Python
• Deep Learning: TensorFlow, Keras
• ML: Scikit-Learn
• NLP: spaCy, NLTK
• Recommendation Systems: Surprise
• Data: Pandas, NumPy
• Visualization: Matplotlib, Seaborn
• Environment: Poetry, Jupyter Notebook, Google Colab
• Version Control: Git, GitHub