If your organisation is interested in delivering a hands-on session at the CEDS UK conference please contact DSNE@lancaster.ac.uk.

More sessions will be added as they are confirmed.

Introduction to Machine Learning with Stephen Hadad, Met Office, UK. 

Tuesday 5th July, 15.00-17.00

Environmental Sciences are a data rich field, with ever growing data volumes from higher-resolution models and new sensors such as from citizen scientists and internet of things (IoT) objects. This presents both an opportunity to gain new insights and understanding of how our natural environment functions, but also a challenge to be able to fully exploit this resource. Data science and machine learning tools and techniques are an essential tool for the environmental scientist to be able exploit this opportunity and meet this challenge.

This workshop will introduce key topics in data science and machine learning through hands-on examples drawn from weather & climate with real data. The workshop will divided into three themes:

• Data Loading and Handling
• Data Exploration and Visualisation
• Building a Machine Learning Pipeline

The workshop will last 2 hours (including ten minute break) and be delivered as a series of Jupyter notebooks containing practical real-world examples. For each theme, the key topics and topics will be introduced by the presenter, followed by time for participants to work through the example notebook. Participants are expected to bring their own laptop and will be supplied with instructions ahead of the workshop to set up a suitable environment for running the notebooks. The notebooks will be freely available on Github so participants can continue to use this resource after the workshop.

The material for the workshop are now available via the link below (google.doc). That will have details in it on setting up a conda environment, which participants will be expected to do before the workshop. This will be sent to participants the week before the workshop so they can prepare and make the most of the workshop. Participants will also then be able to go through material in slower time afterwards.

https://docs.google.com/document/d/1HL9TsYr3jOx3NRUcgw1RGdSElaWCNnVLT3yYbIxCmuQ/edit?usp=sharing

Stop Clicking and Start Coding – GIS Programming in R with Barry Rowlingson, Senior Research Fellow, Lancaster Medical School

Wednesday 6th July, 13.45-15.45

Is your hand on your mouse more than your keyboard when using your GIS? Do you want to be able to repeat things without inflaming your already painful RSI? Then you need to do your GIS work in a programming language!

Several languages now have the power to easily read, manipulate, and output spatial data, using common low-level libraries of optimised code for spatial operations such as coordinate transforms and vector data operations.

For this workshop I’ll talk a bit about the underlying standards and structures of modern open-source spatial data, and then let you work on some examples using R. I’ll supply a set of vector and raster data with which you can work through some exercises. These will involve the “classic” GIS operations (buffering, intersection, point-in-polygon etc) and also some more complex processes that will involve a little programming.

A little prior experience with R might be useful so you don’t spend the first half hour getting nowhere. Make sure you have R installed on your laptop (see www.rstudio.com for installation of R and the RStudio interface) and then install the packages “sf”, “raster”, and “tmap”. These are all open source and run on Windows, Mac and Linux.

By the end of the two hours, you should be able to create a short reproducible analysis report without reaching for your mouse!