Machine Learning for Geospatial Data Science

Learning Outcomes

Learning Outcomes

Having successfully completed this module you will be able to:

• Recognise the role and implications of large language models (e.g., ChatGPT) and apply them responsibly in geosciences [optional].
• Gain hands-on experience with geospatial cloud computing platforms (e.g., Google Earth Engine) to process, manage, and analyse large-scale geospatial datasets, a critical step for geospatial machine learning applications.
• Design and implement – or effectively apply existing – machine & deep learning algorithms to geospatial challenges, including land use classification, air quality, and climate prediction.
• Understand the fundamental theories behind classical machine learning methods (classification and regression) and the two main branches of deep learning (computer vision and natural language processing).

This module combines lectures and practical sessions to provide students with both theoretical knowledge and applicable skills. Part 1: Advanced programming and data science [3 weeks] Week 1: Introductory lecture will cover the fundamentals of Geospatial Data Science, covering: •What is Geospatial Data: Understanding spatial data types, structures, and how they differ from traditional datasets. •Why Geospatial Data is Important: Exploring its role in real-world applications such as climate analysis, urban planning, and environmental monitoring. •Where to Find Geospatial Data: Overview of key sources including satellite imagery and public data portals. •How to Use Geospatial Data: Introduction to basic tools and techniques for processing, analysing, and visualising spatial data. Week 2: Hands-on introduction and refresher on programming language(s) and environments & tools: command line, code editors (e.g., VS Code), IDEs (e.g., Spyder and PyCharm), and Jupyter Notebook/Lab. The session also covers basic Python syntax and provides guidance for students transitioning from R to Python. Week 3: Practical use of geospatial Python modules for managing and processing geospatial data (e.g., satellite imagery), including Shapely, GeoPandas, Rasterio, xarray, and rioxarray. Part 2: Machine/deep learning [4 weeks] Week 4: Introduction to Geospatial AI. This lecture will provide: •an overview of the evolution of artificial intelligence •its growing relevance in the geospatial field •the critical role of high-quality data in building effective models. Week 5: Traditional machine learning (classification): Land classification Week 6: Traditional machine learning (regression): Climate prediction Select one option for Week 7: Week 7a: Deep Learning — Object Recognition, Super Super-Resolution Imaging, or Image Segmentation Week 7b: Generative AI and Large Language Models for Programming Week 7c: Traditional machine learning (advanced): Hybrid modelling Part 3: Cloud computing [4 weeks] Week 8: Introductory lecture on Cloud Geospatial Computing, covering: •What is Cloud Computing? •Importance and Benefits of Cloud-Based Spatial Analysis •Common Cloud Platforms for Geospatial Data Processing Week 9: Explore and become familiar with Cloud Platforms (e.g., Google Earth Engine), focusing on data search and visualisation. Week 10: Practice data processing tasks (e.g., quality control, cloud masking, and data downloads) using Python with APIs of e.g., (Google Earth Engine, Microsoft Planetary Computer, Copernicus Data Space Ecosystem, and Climate Data Store). Week 11: Revisit geospatial machine learning and link it to Cloud Platforms.

The module will include a series of introductory lectures covering data science, programming, machine learning, and cloud computing, with an emphasis on how these areas are interconnected. It will also provide practical sessions held in computing laboratories. Each session will begin with a brief lecture introducing the background of the content being used and highlighting the key concepts addressed in the practical. All course materials – including lecture notes, practical handouts, relevant datasets, and coursework details – will be made available through Blackboard.

Formative

This is how we’ll give you feedback as you are learning. It is not a formal test or exam.

Class practicals

• Assessment Type: Formative
• Feedback: Students will receive regular formative feedback during practical sessions by comparing their work with provided worked examples and discussing any differences between their solutions and the reference code.
• Final Assessment:
• Group Work: No

Summative

This is how we’ll formally assess what you have learned in this module.

Referral

This is how we’ll assess you if you don’t meet the criteria to pass this module.