This repository contains the material (notebooks, data) for the pandas tutorial at EuroScipy 2016. For previous versions of the tutorial (EuroScipy 2015), see the releases page.

To follow this tutorial you need to have the following packages installed:

• Python version 2.6-2.7 or 3.3-3.5
• pandas version 0.18.0 or later: https://pandas.pydata.org/ (previous versions will work for most examples as well)
• numpy version 1.7 or later: https://www.numpy.org/
• matplotlib version 1.3 or later: https://matplotlib.org/
• ipython version 3.x with notebook support, or ipython 4.x combined with jupyter: https://ipython.org
• seaborn (this is used for some plotting, but not necessary to follow the tutorial): https://stanford.edu/~mwaskom/software/seaborn/

I recommend to use the conda environment manager to install all the requirements (you can install miniconda or install the (very large) Anaconda software distribution, found at https://continuum.io/downloads).

Once this is installed, the following command will install all required packages in your Python environment:

conda install pandas jupyter seaborn

But of course, using another distribution (e.g. Enthought Canopy) or pip is good as well, as long as you have the above packages installed.

If you have git installed, you can get the material in this tutorial by cloning this repo:

git clone https://github.com/jorisvandenbossche/pandas-tutorial.git

As an alternative, you can download it as a zip file: https://github.com/jorisvandenbossche/pandas-tutorial/archive/master.zip. I will probably make some changes until the start of the tutorial, so best to download the latest version then (or do a git pull if you are using git).

Two data files are not included in the repo, you can download them from: titles.csv and cast.csv and put them in the /data folder.

To view the content on nbviewer:

• Index
• 01 - Introduction
• 02 - Data structures
• 03 - Indexing and selecting data
• 03b - Some more advanced indexing
• 04 - Groupby operations
• 05 - Time series data
• 06 - Reshaping data