Syllabus | Module 3

Day 1 | Monday, March 3, 2025

Module 3 Introduction (30 min)

• Welcome Back
• Interim Check In
• Where we are in Earth Systems Data Science in the Cloud
• Course Goals and Objectives
• Module Goals and Objectives
• Course Logistics

AI/ML Overview (60 min)

• All of Statistics
• Understanding & Prediction
• Bayesian & Frequentist
• Machine Learning & AI
• Unsupervised, Semi-Supervised, Supervised Machine Learning
• Classification and Regression
• Deep Learning and Not-so-Deep Learning
• Types of data: Image, Text, Gridded, Tabular
• Manual and Automated ML

Demo Day | Parallelization & ML (90 min)

• Machine Learning: The Bad, the Good, and the Awesome!
• Scale: Data Processing, Training, and Inference

Lunch and Learn

• Individual and Team Progress Check In

Model Development Workflow (90 min)

• Goals
• Metrics
• Training and Testing Data
• Preparing your data
  • Information Leakage
  • Order of Operations
• Training
  • Cross validation!
• Optimization
• Fitting
• Out of Sample Performance
• Workflow Evaluation

Team Project Module Goals (30 min)

• Team Project Check-In
• Team Project Goals
• Presentation
• Report

Team Project Work - Data Processing (60 min)

• Data Preprocessing

Day 2 | Tuesday, March 4, 2025

Feature Engineering (60 min)

• Transformation
  • Imputation
  • Actual Transformation
  • One-hot Encoding
  • Outliers
  • Scaling
• Extraction
  • Dimension Reduction
  • Clustering
  • Binning
• Creation
  • Indexing
  • Orthogonal Matching
  • Lagging

Multi-Worker Parallelization | Dask (60 min)

• Starting a Cluster
• Running a Cluster
• Troubleshooting a Cluster
• Shutting Down a Cluster

Building Blocks of Machine Learning (60 min)

• Linear Regression
• Logistic Regression
• K-Means Clustering

Lunch and Learn

• Machine Learning Background Discussion

Machine Learning Metrics (30 min)

• What they mean and how to interpret/implement them
• Classification Metrics
• Regression Metrics
• Bias Variance Tradeoff
• Stats Modeling criteria
• Unbalanced data considerations

Team Project Time

• Feature Engineering

Day 3 | Wednesday, March 5, 2025

The Tool Landscape (30 min)

• SciKit Learn
• Darts
• Deep Learning: Tensorflow, PyTorch, Keras
• Hugging Face
• AutoML Frameworks (Gluon, Canvas)

Serverless Multi-Worker Parallelization (150 min)

• Container Orchestration
• Cloud services
• Servers vs Serverless
• Lambda, Fargate, Lithops, Coiled, Modal

Lunch and Learn

Explainable AI (60 min)

• What is XAI?
• Principles
• Applications
• Techniques

Team Project Work

• Beginning ML

Day 4 | Thursday, March 6, 2025

Cross Validation (60 min)

• Best Practices
• Strategies
• Validating across Space and Time

Team Project Development (120 min)

• Report development
• Presentation development

Lunch and Learn

• Team Project Check In

ML Algorithms and Approaches (90 min)

• Unsupervised
• Semi-supervised
• Supervised

Team Presentation Training (90 min)

• Clear Communication
• Body Language
• Story Telling
• Methods Focus
• Takeaways

Day 5 | Friday, March 7, 2025

Team Presentations | DPD (120 min)

• Capstone Presentations and Feedback.

Module Wrap Up (30 min)

• Closing
• Moving to Model Development
• Interim Period
• Next Steps