# SciTools/cartopy-tutorial

Cartopy tutorial: Around the world in 80 ways
bjlittle Merge pull request #9 from bjlittle/load-exercise-solutions
`tidy loading/showing solutions`
Latest commit 42cb770 Jul 11, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.
appendix Jul 9, 2018
binder Jul 6, 2018
env-exports Jun 18, 2018
resources Jul 9, 2018
static Jul 8, 2018
tutorial Jul 11, 2018
.gitignore Jul 4, 2018
check_env.py Jun 18, 2018
introduction.ipynb Jul 8, 2018
requirements.txt Jun 14, 2018

# SciPy 2018

Date: 8am on Monday 9th July, 2018

## Abstract

On the 1st of October 1872, Jules Verne’s character Phileas Fogg embarks on an adventure to circumnavigate the world in just 80 days. Travelling exclusively by sea and rail, Fogg voyages from the misty alleys of Victorian London to the exotic subcontinent and the Wild West in a race against the clock. Despite fastidiously counting each sunrise, he famously neglects the effect of crossing the dateline and thus arrives on a different day to the one he calculated.

The mistreatment of the dateline is far from unique to Phileas Fogg; it is not just the 24hr time discontinuity that we struggle with - the antimeridian is also a longitudinal discontinuity of 360 degrees for coordinate systems centered at the Greenwich Prime meridian. Whilst Cartesian treatment of geospatial data is a reasonable assumption with an appropriately transformed coordinate system and at large scale (small area), the discontinuities of the antimeridian and the poles frequently wreak havoc on our Cartesian tools when we treat larger area (global / small scale) data.

In this tutorial, join Phileas Elson on our own epic adventure by tracking Fogg’s travel log with careful non-Cartesian treatment as we go. Along the way, we will discover how many of the python libraries we may already know and love can be used in conjunction with cartopy to provide a powerful suite of open source geospatial tools.

We will cover many of the core concepts in cartopy, including topics such as: coordinate reference systems and projections; the matplotlib interface; geospatial image, data (NetCDF) and raster treatment; as well as geometry predicates and transformations. The knowledge developed from this tutorial will be applicable to a broad range of geospatial analyses (both raster and vector), and will provide hands-on experience of tools such as numpy, proj.4, matplotlib and shapely to name a few.

The tutorial will be applicable to a broad range of experiences, but familiarity with numpy and matplotlib will be essential to enable progress through the exercises. A prior awareness of tools such as shapely, proj.4, geopandas, and xarray/Iris will help, though are by no means essential. The tutorial will be made up of a number of exercises with each being designed to accommodate a broad range of expertise.

# Installation

Please see the dependency requirements for a list of packages needed to participate in this tutorial.

Of significant note is that this tutorial is only designed for Python 3.6 or greater. Whilst most of the tools are available for earlier versions of Python, the tutorial itself isn't (and will never be).

## Conda

Conda is the recommended tool for installing the dependencies of this tutorial. Users of conda or Anaconda can install the dependencies into a clean environment with:

``````conda create --file=requirements.txt --channel=conda-forge --name=scipy-cartopy-tutorial
``````

There are a number of large packages in the `requirements.txt`, and this could take a few minutes to download.

Once installed, the environment can be activated as usual with:

``````conda activate scipy-cartopy-tutorial
``````

Or for users of older versions of conda `source activate scipy-cartopy-tutorial`.

## Pip

`pip` is not recommended for installing this tutorial's dependencies. This is mostly because there are a number of non-trivial (non-python) packages that need to be installed. These are the kinds of packages that a tool like conda was designed for.

## Conda environment file

The conda environment under which the tutorial was originally written is available in `env-exports/\${platform}.txt`. This may be useful in ascertaining specific dependency versions were used when developing.

These were produced with `conda list --explicit > env-exports/\${platform}.txt`.

To create a conda environment based on one of the appropriate conda environment files:

``````conda create --name scipy-cartopy-tutorial --file env-exports/\${platform}.txt
``````

# Checking the tutorial installation

First, if you have not already done so, activate the tutorial conda environment:

``````conda activate scipy-cartopy-tutorial
``````

Or for users of older versions of conda `source activate scipy-cartopy-tutorial`.

Now, run the `check_env.py` script in the root of this repository. If all the required packages say `[ OK ]`, then you're ready to start your journey around the world!

# Running the tutorial

Once you have installed and checked the necessary dependencies, run `jupyter notebook` from the root of this repository, and start with `introduction.ipynb`.