The PyData Toolbox
The numerical programming is among the fastest growing areas of application for
Python. The recent explosion of domain-specific tools for scientific computing
in Python can be daunting, but the vast majority of these libraries are built
on a small core of foundational libraries. Understanding these libraries -- how
they work, how they're used, and what problems they aim to solve -- is an
invaluable tool for effectively navigating the PyData ecosystem.
The primary goal of this talk is to provide an introduction to two of these
core libraries: Numpy and Pandas. We focus in particular on motivating the
design of numpy's array class, which serves as the foundational data
structure for numerical computing in Python.
First presented at Boston Algorithmic Trading on Tuesday, Aug 1st, 2017.
Video of the presentation: https://youtu.be/YAHZa8xZWBU.
Running the Presentation
This talk was delivered using Damian Avila's excellent
RISE extension for the Jupyter
Notebook, which allows users to convert a live, executable notebook into a
reveal.js presentation. Assuming you have the necessary system dependencies
(i.e. C and Fortran compiler toolchains), the run.sh script included in the
root of the repo should be sufficient to install, configure, and run the talk.
$ git clone git@github.com:ssanderson/pydata-toolbox.git
$ cd pydata-toolbox
$ ./run.sh
run.sh will create a virtualenv named venv with all necessary dependencies
in the root directory of this project. It will then start an instance of the
Jupyter Notebook server with the RISE extension installed and enabled.