An introduction to scientific computing with Python
by Conor Lawless
Workshop notes: http://bit.ly/NCLpy
Workshop objectives:
Overall the goal is to be in a good position to continue learning to program with Python independently. Specific steps for the workshop will be to:
- Install Python
- Write some very simple Python scripts
- Run Python scripts in a few different ways
- Understand basic programming commands
- Find and install libraries for extending Python
Learning by doing is probably the best way to advance with programming, and it's fun and satisfying. There are lots of online resources for continued learning.
- Why is programming important for bioscience?
- Why Python specifically?
- Why not R?
- What about other programming languages?
- Some practical examples
Why should I learn Python?
Why is programming important for bioscience?
- Much biological data are electronic (e.g. microscopy images, medical scans, genome sequences, log files)
- Digital data are best stored and analysed by computer
- Modern computers: incredibly powerful, often underutilised by biologists
- Learning a programming language allows humans to command computers
- Digital literacy increasingly important for efficient science (e.g. py4Science at Vienna Biocenter)
- Independence from commercial software developers
- Even basic familiarity increases productivity of multi-disciplinary research
Why Python specifically?
- Free to install on any computer, at no cost (independence)
- Python is multi-platform (Windows, OSX, Linux)
- Skills developed are fully transferrable
- Open to other researchers (reproducible science)
- Python syntax is very clear
- Python is easily extensible
- Writing Python code is fast
Open, easy to learn, easy to use
Why not R?
- R is also awesome
- There are some great R courses at Newcastle
- R is also free, open source and extensible
- However, R syntax is ugly: difficult to read and write
- Developed for statistical programming (for which it is excellent)
- R was not developed for general-purpose computing
- I use both, swapping between Python and R
Relatively difficult to learn. Easier to learn to program with Python first.
What about other programming languages?
- Some other languages are more powerful: generate programs that run more quickly (e.g. C++, C, Java)
- Syntax is more verbose and more difficult to learn
- C++, C, Java: easier to make mistakes, more difficult to debug
- I use C++, but increasingly less often, as development is much slower than in Python
- Mathematica and Matlab: tools for applied maths & data visualisation however they require costly licenses
- Perl is an older open source language but notoriously difficult to read and learn
Powerful & fast but difficult or expensive. Easier to learn to program with Python first.
Two example Python scripts for the lab
- Manually counting Colony Forming Units (CFUs) on agar plates (code)
- Resizing and joining digital images (code)
