Ipython — Jupyter — Scientific Library
Howto Install and Start Using
16 Oct 2015
ipython is an excellent and easy way to implement rapidly and nicely some of the techniques we learned during the lecture. The main advantages of Python are the following:
- Beautiful and very natural syntax, easy to learn;
- Huge and increasing community – both in science and industry;
- Loads of libraries for scientific computing: numpy, scipy, matplotlib, pandas, etc.;
- Language of reference for data analysis, computing, financial engineering and statistics;
- Good balance between low and high level programing language: Higher than C++ but possibility to implement C through Cython;
- Excellent cross platform interfaces: ipython notebook, Sage, etc.;
- Produce beautiful pictures, and we all love pictures: matplotlib, seaborn, etc.;
- Free as in “Free Beer” or “Free Speech”: Open source! You do not have to pay for it and can contribute to the libraries. Central for science.
I won’t discuss in details why, some did it better than I would ever do there or there. There are many other alternatives, but to my mind none of them does have all the advantages listed above yet – perhaps in the next future with Julia for instance that also can run in ipython notebook. In any cases, this is really by far better than Matlab – we hate the dinosaur Matlab with its painful and expensive closed source business – see here or here.
At the end of this page, you can find the explanation how to install it.
Some Tutorials, Useful Links, Additional Information
If you are new to Python, I advise you to follow the online course at codeacademy. After registration you can choose to follow the python course.
After that, try playing around with tutorial ipython notebooks that you can download on your computer and use. For instance Patrick Varilly produced a series of notebooks to learn how to manipulate data, use the libraries and plot things. Learning by doing is the better way.
You will make a lot of use of numpy and scipy as well as matplotlib, the library for scientific programing. Here also, learn by the examples – in particular for plots there are thousands of them. However, if you need exact documentation about one or the other function, look at numpy reference,scipy reference or matplotlib reference.
Couple of tutorials:
- Numpy a little bit outdated
- Numpy/Scipy/Matplotlib
More advanced courses:
- Official Python 3.4 tutorial
- Learn Python the Hard Way
- Python course by google excellent but you will need vpn to access
- Otherwise look at online courses provided by universities – I personally do not like to follow a course on video in particular programing
Finally, there are many ipython notebooks online where you can learn about different topic. Here is a – not fully up to date – list of notebooks classified by topics.
BE AWARE: that there are some small differences between Python 2 and 3, the main one being
print “Hello” # for python 2
print(“Hello”) # for python 3
So if you download a notebook, it may happen that it is not fully compatible with your version so you will have to debug a little bit.
Install
The easiest way to get Python together with all the useful libraries as well as the GUI interfaces is to install Anaconda
You’re a Micro$$oft Windows user:
- Download the Anaconda installer for windows by choosing Python 3.4 and the version for your computer 32-bits or 64-bits.
- Double click on the downloaded file, follow instructions
-
Once installed, launch the program anaconda command line and type
pip install seaborn
- You can now run the program Ipyhon notebook that will open ipython in your browser
You’re a Mac O$$X user:
- Download the Anaconda installer for linux by choosing Python 3.4 and the version for your computer 32-bits or 64-bits.
- Double click on the downloaded file, follow instructions
-
Once installed, launch the program anaconda command line and type
pip install seaborn
- You can now run the program Ipyhon notebook that will open ipython in your browser
You’re a linux user (Ubuntu distribution \(\geq\) 14.04)
You can also install anaconda the same way as above. However you can also set up a virtual environment to isolate what you are doing from the rest of the system as follows:
copy paste in terminal (ctrl+alt+t)
sudo apt-get install -y python3-dev g++ libblas-dev liblapack-dev gfortran libfreetype6-dev libpng-dev
Then install globally virtual environment by typing
sudo pip install -y virtualenv virtualenvwrapper
and create a virtualenv in the directory ipython where you will be working
virtualenv –no-site-packages ipython
Activate the virtualenv
cd ipython
source ./bin/activate
And then install what is needed
pip install numpy scipy matplotlib ipython[all] patsy pandas statmodels scikit-learn seaborn
That’s pretty much it, close the terminal.
Inside the directory IPyhton, you can create a directory _notebook where you will store and organize your notebooks.
To run ipython from this directory you go in a terminal, activate the virtual environment, move to the directory _notebook, and run ipython notebook:
cd ipython
source ./bin/activate
cd _notebook
ipython notebook