Data Munging With Python and pandas

Notebook for an 18F tech talk about solving common data munging challenges with Pandas.

Overview

Nothing fancy here...this talk is a brief tour of the Pandas library and its role in basic (smallish) data exploration and munging tasks.

Goals:

• provide a quick overview for potential new Pandas users
• give a basic refresher (with links to resources for more advanced tasks) to people who already use Pandas

Setup

To run the code samples in the Jupyer notebook interactively:

1. Install the miniconda Python package manager

2. From a terminal, clone this project repository to your local machine: [email protected]:bsweger/pandas-munging.git

3. If you don't have a GitHub account and want to get a read-only version of the code, use this command instead: git://github.com/bsweger/pandas-munging.git

4. Change to the project directory:

    cd pandas-munging
   

5. Install Python and dependencies into a conda virtual environment called pandas-munging.

    conda env create
   

6. Activate the conda environment:

    source activate pandas-munging
   

7. Start up the Jupyter notebook server:

    jupyter notebook
   

8. The notebook startup process should have opened a Jupyter web page at localhost:8888 (if not, point your browser there).

9. On the Files tab of the Jupyter home page, click pandas_data_munging.ipynb.

10. The Jupyter notebook with the slides and code from the presentation should open.