README.md

CS 404/504 – Special Topics: Python Programming for Data Science

University of Idaho - Department of Computer Science

Instructor: Alex Vakanski (vakanski@uidaho.edu)

Teaching Assistant: Longze Li (li8975@vandals.uidaho.edu)

Semester: Fall 2023 (August 21 – December 15)

Course Syllabus

Course website: https://fall-2023-python-programming-for-data-science.readthedocs.io/en/latest/

GitHub repository: https://github.com/avakanski/Fall-2023-Python-Programming-for-Data-Science/blob/main/README.md

Lectures:

• Lecture 1 - A Short History and Current State of Artificial Intelligence

Theme 1: Python Programming

• Lecture 2 - Data Types in Python
• Lecture 3 - Statements, Files
• Lecture 4 - Functions, Iterators, Generators
• Lecture 5 - Object-Oriented Programming
• Lecture 6 - Exceptions, Modules and Packages
• Tutorial 1 - Jupyter Notebooks
• Tutorial 2 - Terminal and Command Line
• Tutorial 3 - Python IDEs, VS Codes

Theme 2: Data Engineering Pipelines

• Lecture 7 - NumPy for Array Operations
• Lecture 8 - Data Manipulation with Pandas
• Lecture 9 - Data Visualization with Matplotlib
• Lecture 10 - Databases and SQL
• Lecture 11 - Data Exploration and Preprocessing
• Lecture 12 - Data Visualization with Seaborn
• Tutorial 4 - Virtual Environments
• Tutorial 5 - Web Scraping
• Tutorial 6 - Google Colab

Theme 3: Model Engineering Pipelines

• Lecture 13 - Scikit-Learn Library for Data Sceince
• Lecture 14 - Ensemble Methods
• Lecture 15 - Artificial Neural Networks
• Lecture 16 - Convolutional Neural Networks with Keras and TensorFlow
• Lecture 17 - Model Selection, Hyperparameter Tuning
• Lecture 18 - Neural Networks with PyTorch
• Lecture 19 - Natural Language Processing
• Lecture 20 - Transformer Networks
• Lecture 21 - NLP with Hugging Face
• Lecture 22 - Diffusion Models for Text-to-Image Generation
• Lecture 23 - Large Language Models
• Tutorial 8 - TensorFlow
• Tutorial 9 - PyTorch
• Tutorial 10 - TensorFlow Datasets
• Tutorial 11 - Experiment Monitoring with CometML
• Tutorial 12 - GitHub

Theme 4: Model Deployment Pipelines

• Lecture 24 - Introduction to Data Science Operations (DSOps)
• Lecture 25 - Deploying Projects as Web Applications
• Lecture 26 - Deploying Projects to the Cloud
• Lecture 27 - Reproducible Data Science Projects
• Tutorial 13 - GitHub Actions

Course Description

The course is designed to introduce students to Python tools and libraries that are commonly used by organizations for managing the various phases in the life cycle of data science projects. The content is divided into four main themes. The first theme reviews the fundamentals of Python programming. The second theme focuses on data engineering and explores Python tools for data collection, exploration, and visualization. The next theme covers model engineering and includes topics related to model design, selection, and evaluation for image processing, natural language processing, and time series analysis. The last theme introduces Data Science Operations (DSOps) and encompasses techniques for model serving, performance monitoring, diagnosis, and reproducibility of data science projects deployed in production. Throughout the course, students will gain hands-on experience with various Python libraries for data science workflow management.

Textbooks

1. Joel Grus, “Data Science from Scratch: First Principles with Python,” 2nd Edition, O'Reilly Media, 2019, ISBN: 9781492041139.
2. Chip Huyen, “Designing Machine Learning Systems,” O'Reilly Media, 2022, ISBN: 9781098107963.

Learning Outcomes

Upon the completion of the course, the students should demonstrate the ability to:

1. Attain proficiency with commonly used Python frameworks for managing the life cycle of data science projects.
2. Develop pipelines for integrating data from multiple sources, designing predictive models, and deploying the models.
3. Apply Python tools for data collection, analysis, and visualization, such as NumPy, Pandas, Matplotlib, and Seaborn, to real-world datasets.
4. Implement machine learning algorithms for image processing, natural language processing, and time series analysis using Python-based frameworks, such as Scikit-Learn, Keras, TensorFlow, and PyTorch.
5. Understand the principles of model selection and evaluation, including hyperparameter tuning, cross-validation, and regularization.
6. Understand the primary characteristics of current Python libraries for deployment, continuous integration, and monitoring of data science projects.
7. Deploy data science projects as web applications using Flask, and to cloud servers using Microsoft’s Azure platform.

Prerequisites

The course requires to have basic programming skills in Python. While having knowledge of data science methods would be advantageous, it is not mandatory.

Grading

Student assessment will be based on 6 homework assignments (worth 60 pts), 3 quizzes (worth 30 marks), and class participation and engagement (worth 10 marks).