Stats 507, Data Science in Python

About readings

Reading assignments are listed below by week. Generally readings will prepare you for content to be discussed in lecture that week, though there will not always be a perfect correspondence. Sometimes readings may cover older material in more depth or may lay some groundwork for future discussions. I generally expect students will have completed some or all of a week’s readings prior to that Tuesday’s lecture.

Some readings will be marked “recommended” and or “optional” with “optional” readings intended for interested students only. Both “recommended” and “optional” texts are okay to skip for those struggling to keep up.

Some readings may also be labeled “okay to skim” meaning you should familiarize yourself with what’s there, read what’s helpful to you, and refer back to read in more detail as needed.

There is a close alliance between the Severance and Downey texts and generally it will be sufficient to do the readings from one or the other. There is some redundancy built into these readings and the McKinney readings - pick one to read first and skim the other if needed.

I will generally aim to post readings by Monday evening a week prior to their due date.

Texts

Readings from the following texts will be referred to by the author’s last name.

• Think Python 2, Allen B. Downey.

• Python for Everybody, Charles R. Severance.

• Python for Data Analysis, Wes McKinney. (Accessible online to UM students.)

Reading List

• Week of August 31
  • McKinney: Preliminaries (okay to skim) & Chapter 2 - Python Language Basics, IPython, and Jupyter Notebooks
  • Jupyter Notebook Documentation, read all of “The Jupyter Notebook” and okay to skim the rest.
  • Downey chapters 1-3 or Severance chapters 1-2, and 4.
• Week of September 7
  • Downey chapters 5-8 or Severance chapters 3, 5, and 6.
  • Downey chapters 10-12 or Severance chapters 8-10.
  • McKinney chapter 3 Built-in Data Structures, Functions, and Files.
  • (Recommended), Downey chapters 9 and 13.
• Week of September 14
  • From Markdown Guide “What is Markdown?”, “Why Use Markdown?”, “How Does it Work?”, “Flavors of Markdown” and all of Basic Syntax.
  • Github Flavored Markdown through “Tables”, okay to skim.
  • The Jupytext readme and (okay to skim) documentation, focus on the light and (optional) markdown formats.
  • McKinney chapters 4-5, “NumPy Basics: Arrays and Vectorized Computation” and “Getting Started with Pandas”.
• Week of September 21
  • Pandas User Guide: Read “10 minutes to Pandas” and (optionally) skim the rest. Additional parts will be assigned in future weeks.
  • McKinney chapters 7 “Data Cleaning And Preparation” and chapter 8 “Data Wrangling: Join, Combine, And Reshape”.
    • Okay to skim these sections/headings: 7.1, 7.2 “Discretization and Binning”, 7.3 “Regular Expressions”.
• Week of September 28
  • McKinney chapter 9 “Plotting and Visualization”, okay to skim sections on seaborne, but read more carefully sections on pandas methods.
    
  • PyPlot tutorial (okay to skim)
  • Chart Visualization in the pandas docs (okay to skim)
  • Categorical data type in the Pandas User Guide (Added 9/23.)
• Week of October 5
  • McKinney chapter 10 “Data Aggregation and Group Operations”. (Will be discussed in class week of 9/28.)
  • Bootstrapping and Subsampling (optional).
• Week of October 12
  • statsmodels User Guide
    • Background,
    • Linear Regression,
    • Generalized Linear Models.
• Week of October 19
  • Fall Break - everything here is okay to skim.
  • McKinney 7.3 “Regular Expressions” (previously skimmed)
  • Severance Chapter 11 “Regular Expressions” (optional)
  • Python’s Regular Expression HOWTO (skim)
  • RegexOne
• Week of October 26
  • Atlassian’s git tutorials, specifically:
    • “What is version control?”
    • “What is git?”
    • Everything under “Getting Started” (okay to skim).
  • Data Science at the Command Line by Jeroen Janssens, here.
    • Introduction and Chapter 4 (both optional)
    • It is not expected that you install the Docker image.
  • How to Write a Git Commit Message (optional)
  • The GitHub learning lab (optional)
  • Visualizing git commands (optional)
  • A quick explanation of merge conflicts (optional): https://www.youtube.com/watch?v=sfT0WrChMrM
• Week of November 2
  • Elements of Statistical Learning:
    • Model Assessment and Selection (7.1-7.3, 7.10)
  • Scikit-learn User Guide:
    • 1.1.1 OLS, 1.1.2 Ridge Regression, 1.1.3 Lasso, and 1.1.5 Elastic Net
• Week of November 9
  • Elements of Statistical Learning:
    • Boosting and Additive Trees (10 optional, 10.1, 10.2, 10.9, 10.12 are highly recommended),
    • Random Forests (15, optional).
  • Scikit-learn User Guide:
    • 1.11.2 Forest of Randomized Trees
    • 1.11.4 Gradient Tree Boosting
  • XGBoost Python Package (okay to skim)
• Week of November 16
  • multiprocessing docs (optional)
  • GreatLakes User Guide read “1.3 GETTING STARTED (COMMAND LINE)” and skim rest.
  • Slurm User Guide