3-Day Subsurface Maching Learning, Spatial Data Analytics, and Geostatistics Course

1-Day Spatial Data Analytics Course + 2-Day Machine Learning Course

This repository contains files related to a training class dated 12/08/2025.

Instructor: Michael Pyrcz, Professor, The University of Texas at Austin

DIRECT Consortium | daytum | Twitter | YouTube | LinkedIn | Webpage | Geostats Book | Machine Learning e-Book | Geostatistics e-Book | GoogleScholar

Instructor: John T. Foster, Professor, The University of Texas at Austin

DIRECT Consortium | daytum | Twitter | YouTube | Webpage | High perforance Computing e-Book | Introduction to Python e-Book | GoogleScholar

Course Summary

Building from fundamental probability and statistics, we cover entire spatial data analytics, geostatistics and machine learning best practice workflows from data preparation through to decision making. We will accomplish this with,

• Interactive lectures / discussion to cover the basic concepts

• Demonstrations of methods and workflows in Python

• Hands-on experiential learning with well-documented workflows for accessibility

Course Objectives

Spatial data analytics and geostatistics for building spatial prediction and uncertainty models.

You will learn:

• spatial data debiasing

• quantification and modeling of spatial continuity / correlation

• spatial estimation with uncertainty

• spatial simulation for subsurface resource forecasting

• checking spatial models

• decision making with spatial uncertainty models

• inferential machine learning

• predictive machine learning

---

Course Schedule

The following tables include the,

• approximate times - the nominal schedule. Note, we are learning and not schedule-driven; therefore the course delivery will adjust for the needs of the class.

• topics - general topic covered.

• objectives - major objective of the session as the new knowledge or skill set.

• lecture notes - link to lecture notes in PDF. In some cases I have included notes directly from my UT Austin courses to offer greater coverage.

• demos - well-documented workflow demonstrating the theory and best practice from the course notes.

• interactive - interactive Python dashboards to demonstrate a concept.

• e-book - link to the associated chapter from Dr. Pyrcz's free, online e-books.

• lecture - link to the associated recorded lecture from Dr. Pyrcz's YouTube channel.

Spatial Data Analytics and Geostatistics 1-day Short Course

Day	Time	Topic	Objective	Notes	Demo	Interactive	e-book	Lecture
Day 1	8:00 AM - 8:30 AM	Course Overview	Walk-through of the course plan, goals, methods and introductions	Overview
	8:30 AM - 9:00 AM	Introduction	Data analytics and geostatistics concepts	Introduction			Book	Lecture
	9:00 AM - 10:00 AM	Data Analytics	Multivariate statistical methods to support spatial modeling	Notes	Demo	Dashboard1 Dashboard2	Book	Lecture
	10:00 AM - 11:00 AM	Spatial Continuity Calculation	Measuring spatial continuity with experimental variograms	Notes	Demo	Dashboard	Book	Lecture
	11:00 AM - 12:00 Noon	Spatial Continuity Modeling	Variogram modeling for quantifying spatial continuity	Notes	Demo	Dashboard	Book	Lecture
	12:00 noon - 1:00 PM	Lunch Break
	1:00 PM - 2:00 PM	Spatial Estimation	Spatial estimators, theory and applications with kriging	Notes	Demo	Dashboard	Book	Lecture1 Lecture2
	2:00 PM - 2:30 PM	Simulation and Uncertainty Modeling	Stochastic realizations for uncertainty modeling	Notes	Demo	Dashboard	Book	Lecture
	2:30 PM - 3:00 PM	Advanced Simulation (Optional)	Cosimulation for bivariate simulation models	Notes
			Indicator simulation	Notes	Demo		Book	Lecture
			Multiple point and object-based simulation	Notes
	3:00 PM - 4:00 PM	Model Checking	Essential quality assurance methods for spatial, geostatistical models	Notes	Demo		Book	Lecture
	4:00 PM - 5:00 PM	Decision Making with Uncertainty	Making the best decision in the precense of uncertainty	Notes		Dashboard	Book	Lecture

Machine Learning 2-day Short Course

Day	Time	Topic	Objective	Notes	Demo	Interactive	e-book	Lecture
Day 2	8:00 AM - 8:30 AM	Course Overview	Review schedule only	Overview
	8:30 AM - 10:00 AM	Probability	Frequentist and Bayesian probability approaches	Notes		Dashboard	Book	Lecture
	10:00 AM - 11:00 PM	Data Preparation	Data debiasing methods to correct for sampling bias	Notes	Demo	Dashboard	Book	Lecture
			Introduction to bootstrap for uncertainty modeling	Notes	Demo	Dashboard	Book	Lecture
	11:00 PM - 11:30 PM	Feature Imputation	Dealing with missing data	Notes			Book
	11:30 noon - 12:30 PM	Lunch Break
	12:30 PM - 1:30 PM	Feature Selection	Working with the fewest most informative features	Notes	Demo		Book	Lecture
	1:30 PM - 2:30 PM	Cluster Analysis	k-means clustering	Notes	Demo		Book	Lecture
	2:30 PM - 3:30 PM	Advanced Cluster Analysis	Density-based and spectral clustering	Notes	Demo	Dashboard	Book	Lecture
	3:30 PM - 5:00 PM	Dimensionality Reduction	Principal components analysis	Notes	Demo	Dashboard1 Dashboard2	Book	Lecture
Day 3	8:00 AM - 9:00 AM	Predictive Machine Learning	Concepts and workflows for predictive machine learning	Notes		Dashboard	Book	Lecture
	9:00 AM - 9:30 AM	Linear Regression	Start with simple linear prediction models	Notes		Dashboard	Book	Lecture
	9:30 AM - 10:00 AM	k-Nearest Neighbors	Lazy learning with a mapping analogy	Notes	Demo		Book	Lecture
	10:00 AM - 11:00 AM	Näive Bayes	Bayesian classification model	Notes	Demo		Book	Lecture
	11:00 AM - 12:00 noon	Decision Tree	Simple model that extends to powerful ensemble methods	Notes	Demo	Dashboard	Book	Lecture
	12:00 noon - 1:00 PM	Lunch Break
	1:00 PM - 2:00 PM	Bagging and Random Forest	Averaging over trees to reduce model variance	Notes	Demo		Book	Lecture
	2:00 PM - 3:00 PM	Gradient Boosting	Additive weak learners to avoid overfit	Notes			Book	Lecture
	3:00 PM - 4:00 PM	Neural Networks	Powerful deep learning methods	Notes	Demo	Dashboard	Book	Lecture
	4:00 PM - 4:30 PM	Conclusions and Wrap-up	Summarize and discuss	Notes

Beyond the Course

We have the following daytum short courses ready that we would love share, including core courses,

1. Introduction to the Python Ecosystem
2. Introduction to Energy Data Science using Python
3. Introduction to Subsurface Machine Learning
4. Machine Learning for Energy Executives
5. Spatial Data Analytics and Geostatistics

And other courses such as,

1. Pandas for Excel Addicts
2. Information and Game Theory
3. Advanced Machine Learning
4. Custom Courses Designed to Meet the Needs of the Company

There is Much More – the building blocks can be reimplemented and expanded to address various other problems, opportunities. There is much more that we could cover,

• Additional Theory

• More Hands-on / Experiential Learning

• Workflow Development

• Basics of Python / R

• Advanced Data Preparation

• Advanced Model QC

• Methods to Integrate More Geoscience and Engineering

• Integration of Machine Learning Spatial Modeling

We are happy to discuss other, advanced courses and custom courses to meet your teams' educational needs to add value at work with data science.

Daytum's courses have been taken by employees at:

                        

© Copyright daytum 2025. All Rights Reserved