Essential Data Science for Petroleum Geoscientists and Engineers - (Remote Instructor-Led Series)

Module 1 - Overview

• What is Data Science - Overview of the course, and an outline of the scope of data science.
• Data Science for E&P - Addressing the role of data science in E&P and an example application to log data quality control and reconstruction using machine learning.

Module 2 - Data Science Toolkit - Notebooks, Visualization, and Communication

• Overview of the data science toolkit.
• Hands-on workshop introducing the toolkit and getting started with Python scripts and notebooks.
• Overview of how to manage and use Python packages. 
• Hands-on workshop on Python packages covering how to install and manage packages, and how to use packages from your Python notebooks
• Overview of data visualization with SandDance.
• Hands-on workshop introducing SandDance for interactive data visualization using a dataset of offshore wells from the UK Continental Shelf.
• Overview of Markdown - a lightweight markup language for adding simple formatting to plain text documents, and documenting Python notebooks.
• Hands-on workshop on Markdown for formatting text documents and annotating Python notebooks.

Module 3 - Python Fundamentals

• Python 101 - Introduction to Python fundamentals including variables, types, statements, expressions, control flow, and functions.
• Hands-on workshop on Python 101.
• Python 102 - More Python fundamentals including data structures, modules, files and folders, JavaScript Object Notation (JSON), serialization.
• Hands-on workshop on Python 102.

Module 4 - Computational Thinking

• Introduction to Computational Thinking - the analytical and logical processes of decomposing a complex task and expressing it in a form that can be performed by a computer.
• Hands-on workshop on Computational Thinking applied to the design and implementation an interactive base map for UK E&P data.

Module 5 - Exploratory Data Analysis

• Exploratory Data Analysis - Introduction to the Exploratory Data Analysis process and key Python packages: pandas for data analysis and plotly for statistical graphics.
• Hands-on workshop on exploratory data analysis - reading data into pandas data frames, handling dates, merging datasets, creating statistical graphics figures with plotly, exporting figures.
• Statistical Graphics - Demonstration of a gallery of statistical graphics samples 
• Descriptive Statistics - Introduction to univariate and multivariate statistics.

Module 6 - Exploring E&P Data

• Well header data - Introduction to handling well header data (surface location and attributes) using the pandas and plotly packages.
• Hands-on workshop on well header data - including import, data cleaning, date handling, posting well data on cultural/satellite base map and visualizing historical trends.
• Production data - Introduction to handling field production data using the pandas and plotly packages.
• Hands-on workshop on field production data - including import, data cleaning, date handling, queries, visualizing hierarchical and  time series data. 
• Well log data - Introduction to handling wireline logs from LAS files using the lasio, pandas, and plotly packages. 
• Hands-on workshop on well log and tops data - including LAS (.las) file import, merging tops, and data visualization.
• Seismic data - Introduction to handling seismic SEG-Y data using the segyio, and plotly packages.
• Hands-on workshop on seismic data - including SEG-Y (.segy) file import, extracting binary and trace headers, visualizing seismic trace data, and calculating seismic attributes

Module 7 - Geospatial Data

• Coordinate reference systems - Introduction to the geographic and projected coordinate systems, defining a coordinate reference system from EPSG codes, offsets between coordinate reference systems, and transforming positions between reference systems.
• Hands-on workshop on coordinate reference systems.

Module 8 - Machine Learning Fundamentals 

• Machine Learning - introduction to the fundamentals of machine learning including background concepts from probability theory, the different types of machine learning, and the basic workflow to build and evaluate models from data. 
• Python scikit-learn - introduction to the Python scikit-learn package for machine learning, including a demonstration of typical pipelines and workflows. 
• Supervised learning with regression - introduction to regression including traditional linear regression, random forest regression, hyperparameter optimization, and performance evaluation.
• Hands-on workshop on regression for reconstructing wireline logs using random forests.
• Unsupervised Learning - introduction to unsupervised learning for dimensionality reduction, clustering and outlier detection. 
• Hands-on workshop on unsupervised learning for outlier detection and clustering of wireline logs.
• Explainable Machine Learning - introduction to explainable machine learning: techniques for looking inside the so-called black box models of machine learning to understand why particular predictions are made and which variables are important.