While working at Maersk I went to PyData 2019. I found two presentations very useful and I have used learnings from both in my day job.

Testing for Data Science

Testing for Data Science about the use of pytest and unittest for data science. Some of the topics were:

• Simple tests with pytest.
• Parameterized tests with pytest.
• Mocking with unittest, seems great for mocking a rest API.

See more in the repo from the talk in [1].

Maintainable Code for Data Science

Maintainable Code for Data Science about the use of scikit-learn for data science.

The presentation was useful because it showed some practical examples on how one can organize models using the scikit-learn framework. Some of the topics were:

• The Transformer and Estimator abstraction in scikit-learn.
• How to create custom versions of a Transformer and Estimator.
• An example of how to organize a Python module for model building.

See more in the repo from the talk in [2].

References

[1] Testing for Data Science, Github Repo

[2] Maintainable Code for Data Science, Github Repo

Comments

Feel free to comment here below. A Github account is required.