Python for Atmosphere and Ocean Scientists
Daniel Wheeler
NOAA (virtual), 12/11/2020
Operations
- Software -- Jupyter, AOS Python stack
- I'm assuming you have a little experience with Python
- Please ask me to slow down in the chat interface
- What are we doing?
- Organize your software -- Conda environments
- Using Xarray with NetCDF files
- Vectorization (not writing loops in Python)
- Anything else that you'd like to learn
- Present your own data
- Trying to read and manipulate your own data
Study Material
-
SWC, heavily Pandas based
-
http://swcarpentry.github.io/python-novice-gapminder/
-
-
Python like you mean it
-
https://www.pythonlikeyoumeanit.com/module_2.html
-
-
Automate the boring stuff
-
https://automatetheboringstuff.com/
-
Why learn Python?
-
It's a programming language - doesn't matter
-
Alternatives: Matlab, Mathematica, Perl, Ruby, Julia
-
Widely used in science
-
Widely used outside of science (web platforms, data science)
-
Number 3 on TIOBE index (Java, C, Python), https://www.tiobe.com/tiobe-index/
-
What else: huge active community, many scientific libraries, high level, open
-
Some drawbacks (types, reproducibility)
-
What do your friends use?
AOS Stack
Python software is like the wild west compared with Matlab
Python Environments
- What happens if we only use the "system" Python?
- We need many different Python environments
- In the beginning there was pip and it was ok
- How do we share environments?
- 5-10 years ago Conda became popular
- create any number of new environments
- other language support (Ruby, R, Javascript, C extensions)
- Conda has problems
- There are other choices (e.g. Nix)