Develop practical machine learning skills to address complex environmental data challenges.
Location:
Interactive online course using MS Teams
Lancaster and Edinburgh, as indicated by learner demand
Cost:
Students: £449
Professionals: £499
The above prices are the Early Bird discount rates (thereafter £70 more.)
Date:
Spring 2025
Express your interest here, so we can fix a course date!
Short course description:
This interactive two-day course over four mornings will empower you with critical machine learning knowledge and skills. It will give a strong foundation in two widely used machine learning algorithms: random forests and support vector machines. You will join hands-on coding sessions in Python, using real-world environmental datasets such as water chemistry measurements from Loch Leven. You will learn how to explore and answer relevant environmental questions.
The course emphasizes practical learning, encouraging you to apply new skills to your own datasets. By the end of the course, you will be confident in your understanding of the core mathematical concepts. You will be able to implement these algorithms to solve real environmental problems. We encourage learners to apply the learning from the morning sessions in their own time with their own datasets.
Learning outcome:
By the end of the course, you will have a good conceptual understanding of the mathematics of two supervised machine learning algorithms. You will be able to write Python code to apply these algorithms to solve environmental problems.
Objectives:
- Appreciate how machine learning differs from statistical modelling, and the difference between supervised and unsupervised machine learning.
- Learn how decision trees work and how to implement a simple decision tree classifier in Python.
- Understand ensembling and how a random forest classifier works, as well as how to optimise a random forest classifier on environmental data.
- Study loss functions and the gradient descent algorithm, apply this knowledge to build a logistic regression algorithm in Python from scratch.
- Explore Support Vector Machines (SVMs) as a powerful algorithm for solving classification and regression problems. Write code to implement an SVM classifier.
Target audience:
- MSc/ PhD students and academics
- Environmental consultants
- Research software engineers
Level:
Intermediate
You will have Python already installed and work with it at least occasionally to read and write basic code. A basic understanding of Maths (basics of vectors, matrices and calculus) is a pre-requisite for this course.
Places:
Max 20 learners per event.
Hardware and software requirements:
You will need a laptop with > 4 GB RAM and a 2.4 GHz CPU or higher. No GPU requirement.
An internet connection to access UKCEH DataLabs, where we will host the course materials.
Course leaders:
Ezra Kitson, Environmental Data Scientist, UKCEH
Ezra has ten years’ experience analysing environmental data with statistical and machine learning methods. During his PhD at the University of Edinburgh, he worked on machine learning analysis of high-resolution mass spectrometry data and as part of his role as an Environmental Data Scientist has applied machine learning to help solve a range of geospatial and hydrological data analysis problems. Ezra is passionate about Python programming and enjoys studying and teaching mathematical concepts to people with all ranges of numerical experience.
Alba Gomez Segura, Research Software Engineer, UKCEH
Alba is a Research Software Engineer with a background in environmental science and over seven years of experience working at the intersection of technology and biology. She began her career in bioinformatics and later transitioned to environmental research. Alba joined the UK Centre for Ecology and Hydrology (UKCEH), her work includes integrating machine learning models into research pipelines, supporting real-time data analysis, and managing diverse environmental datasets. She is passionate about bridging the gap between machine learning and environmental science, aiming to provide practical tools and techniques that harness data for ecological research and conservation.