Reproducible Research in Python
summary
A messy Dropbox with 30 data files and seven Python scripts becomes a reproducible pipeline that predicts average life expectancy for ~150 WHO member states using a lasso regression; each dataset is downloaded with pandas read_csv and cleaned by scripted functions that strip characters and coerce numerics; environment is captured with virtualenvs and a pip freeze requirements file; notebook prose lives in Jupyter Notebook and code history is tracked with Git and shared on GitHub (three commits shown).
AI-written summary · about
chapters (10)
- 0:00Intro
- 4:32What did I do
- 7:05Leaving exploratory code
- 8:02Why did we do this
- 9:00What is literate programming
- 10:58Setting up your environment
- 11:43Virtual labs
- 20:28Recap
- 22:06Questions
- 30:18Change Management
related talks
45mPython useRs
02h 14mHands-on Python for Programmers
01h 33mFrom Raw Recruit Scripts to Perfect Python
01h 02mDocker for Data Science: Reproducibility and Deployment
38mBuilding Data Pipelines in Python
01h 16mReproducible machine learning and science with python
52mPython as the multitool of data science
43mBehind the Scenes at Gapminder: A Story about Data and Visualizations