harpolea / scientific_programming

Scientific Programming

This repository contains a collection of notebooks and exercises for the scientific programming course at the 2018 NGCM Summer Academy.

The course covers

• Testing
• Continuous integration
• Code coverage

The notebooks are best viewed on nbviewer.

The notebooks rely on a number of python libraries. These can be installed most easily by first creating a runtime environment:

conda create -n ngcm_sciprog python=3 numpy scipy jupyter matplotlib pytest pytest-cov 

Activate this environment:

source activate ngcm_sciprog

Then launch jupyter:

jupyter notebook

scientific_programming's Issues

PCBI journal figure missing

In 05-Publishing - I think this is the seaborn comparison.

starbucks data missing

In 05-Publishing_code the starbucks.csv file needs downloading from the link. Three options:

1. Expand the description to make it clear they must download before running the code;
2. Use eg requests library to download it automatically;
3. Leave it as is and remember to tell them this at the time.

Plotting data set header rows

For 05-Publishing_code: When I run on my Mac, the dataset corresponding to the observed variable has only one header row, as pandas concats the string in quotes. So I need to set skiprows=1 or I get KeyError: 'No_Smoothing'. This is with the latest conda version of pandas.

Opinionated matplotlib link

http://pbpython.com/effective-matplotlib.html

Worth including in 05?

Commenting figure missing

In 04-Documentation it refers to commenting.jpg which isn't in the repo.