Applied Data Science

This repository contains all the assignments and solutions for the Applied Data Science course by Amir Hesam Salavati, including tasks related to data manipulation, visualization, machine learning, and deep learning. The materials of the course are accessible through this link.

Assignments

Assignment 1: Introduction to Pandas and Jupyter Notebooks

• Familiarization with Python tools for data science, including Colab and Kaggle.
• Practical tasks using datasets to explore the fundamentals of data manipulation.

Assignment 2: Exploratory Data Analysis and Data Cleaning

• Techniques for analyzing datasets, handling missing values, and preparing data for modeling.

• Focus on converting data into usable numerical formats and normalizing values.

  

Assignment 3: Data Visualization

• Practice with a variety of visualization techniques (e.g., pie charts, scatter plots, interactive charts) to uncover insights.

  

Assignment 4: Feature Engineering

• Creation of new features using methods like binning, aggregation, and dimensionality reduction (e.g., PCA).

Assignment 5: Accuracy Measures

• Evaluation of regression and classification models using metrics such as MAE, MSE, Precision, Recall, and F1-Score.

Assignment 6: Regression Methods

• Implementation of linear regression, logistic regression, ridge regression, LASSO, and kernel regression for predictive analysis.

  

Assignment 7: Binary Classification Methods

• Exploration of binary and multiclass classification models, including SVMs, KNN, decision trees, and boosting techniques (e.g., XGBoost, LightGBM).

Assignment 8: Multiclass Classification Methods

• Building and training deep learning models using frameworks like Keras and PyTorch.
• Tasks include binary classification, regression, and time-series analysis.

Assignment 9: Neural Networks

• Application of convolutional neural networks (CNNs) for image classification, transfer learning, and data augmentation.

  

Assignment 10: Deep Neural Network

• Implementation of autoencoders, GANs, and tools like Grad-CAM, SHAP, and LIME to interpret model predictions.

Assignment 11: Convolutional Neural Networks, Transfer Learning, and Data Augmentation

• Techniques to handle imbalanced datasets, including SMOTE, cost-sensitive training, and automation tools like TPOT and AutoSklearn.

  

Assignment 12: Autoencoders, GANs, and Explainable AI

• Autoencoders, GANs, and tools like Grad-CAM, SHAP, and LIME to interpret model predictions.

Assignment 13: Imbalanced Data, Pipelines, and AutoML

• Techniques to handle imbalanced datasets, including SMOTE, cost-sensitive training, and automation tools like TPOT and AutoSklearn.

Contributing

Feel free to submit issues or pull requests if you find bugs or want to improve this repository.

Acknowledgements

This work was completed as part of the Applied Data Science course taught at Tehran Institute for Advanced Studies (TeIAS), Khatam University.

License

This project is open-source and available under the MIT License.