In this repository you will find a collection of Jupyter notebooks containing exercises within the broad area of Environmental Modelling. This is currently run as part of the Department of Earth and Environmental Sciences course on Environmental Modelling, but can be widely used. In all notebooks we present demonstrations of complete code, where exercises are based on either modifying that code or we ask you to change parameters and quantify the impact on a simulation.
If you have not used a Jupyter notebook before, I would reccomend checking out the official webpage. There is also information provided in the PDF associated with each notebook. There are multiple ways you can interact with a Jupyter notebook. The most flexible way is to open and interact with them on your own computer. However you can start an instance of every notebook running in the cloud. We provide details below
Having a Python distribution on your own machine is attractive for a number of reasons, not least gaining familiarity with building projects in your own time. If you havent already, I would reccomend installing the Anaconda distribution. You can download a copy using this link. That page will give you the option to download a version for Windows, Mac or Linux. Download the graphical installer and, typically, accept all options. Once you have installed this, now open a terminal. On Windows, go to the menu of options and find 'Anaconda Prompt' under the Anaconda folder. On a Mac, go to Finder -> Utilities -> Terminal. If on a Mac, when in this terminal when you type:
Python
Do you see the reference to Anaconda? For example, you may see something similar to:
Python 3.7.6 (default, Jan 8 2020, 20:23:39) [MSC v.1916 64 bit (AMD64)] :: Anaconda, Inc. on win32
Now we are going to create a virtual environment to run our notebooks in. Virtal environments are a great way of maintaining a 'work space' that is seperate to your default installation. For example, if you are going to start installing lots of bespoke modules, you may sometimes come across a clash of version numbers which then becomes tricky to maintain. In the worst case scenario, this would require a re-installation of Python. So lets create a virtual environment for our project. You can switch-on and switch-ff these virtual environments from the command line/terminal whenever you need them.
If you are on Windows, go back to the Anaconda prompt. If you are on a Mac or Linux, go back to ther terminal. First we need to clone this repository. We should use Git for this, becuase with Git you can keep pulling updates from this repository. If you do not already have Git installed on your machine, you can get it from the download page. Once you have installed this, at the prompt/terminal type:
This will download the project to the location you are in already. You can change this location before running the above command, or move the folder later. Github also gives you the option to download a ZIP file of the entire project if you cannot, or do not want, to use Git. Once you have the project downloaded, open a command prompt/terminal and navigate to the project folder. We are now going to use the file 'environment.yml' to create a new virtual environment. Run the following command:
conda env create -f environment.yml
You will see a number of packages being downloaded by the conda package manager which is part of the Anaconda distribution. Accept any requests and, when finished, you will see a message that resembles the following:
To activate this environment, use
$ conda activate EnvModelling
To deactivate an active environment, use
$ conda deactivate
These are the commands for switching on/off this new virtual environment. Let's switch it on. Type the following in the command prompt/terminal:
conda activate EnvModelling
In the command prompt, you will see the name (EnvModelling) replace (base). Now we can start an interactive Jupyer notebook session. Still within the project folder, type the following:
jupyter notebook
Can you see the project folders and files? You are good to go! Every time you now want to open the notebooks for this project, open either the Anaconda prompt or Terminal, activate the environment and then run the last command from within the project folder.
If you do not, or cannot, run Python from your own machine we have provided the ability for you to interact with these files using Binder. The Binder project offers an easy place to share computing environments to everyone. It allows users to specify custom environments and share them with a single link. Indeed, if you click the link below this will spin-up an individual session for you. Please bare in mind it can take a while to start, and if idle for a short period these sessions will stop. However you can download your notebook file during the session. Everytime you start a Binder link, it will start from scratch.
Google's Colab Co-laboratory is a great platform for developing machine learning and data-science driven applications on the web. It provides access to free GPU resource (Graphics Processing Units). However it also allows us to run Jupyter notebooks from a Github repository if you have a Google account. If you can register or have an existing Google account, using Google Colab is a really nice experience. It will allow you to save individual files and projects to your Google Drive. We dont cover that here. By clicking on the above link it will take you to a page that presents you with options to load existing files from either your Google Drive or from public repositories. However we can also provide you with a clickable link for running individual notebook files, much like Binder. These are given below and are linked to each notebook file. You will likely find these load much quicker than using Binder. However, you may find any images used in the notebook file that are in the Github repo do not load..but not a huge problem. The links to current notebook files are given below:
- Solution notebook:
- Solution notebook: