rballest/ssi2019

Introduction to Data Science in Python

This repository contains Jupyter notebooks, code samples, and all other materials used in the Intro to Data Science in Python course taught by Tal Yarkoni at the 2019 UT Summer Statistics Institute. Links to new sections or resources will be added below as content is added.

Course overview/links

Inactive links will be added the evening before the following day's class at the latest.

• Test notebook
• Day 1: Data Science in Python
• Day 2: Data Wrangling
• Day 3: Visualizing and Analyzing
• Day 4: Predictive Modeling

Run online

You can run these notebooks online, without installing anything, using Binder! Just click the button below to launch. Note that some cells may fail when running in Binder, and execution will generally be much slower.

Setup

To get the most out of the course, you'll probably want to show up to class with a laptop set up with a working scientific Python environment. The easiest way to accomplish this is to install the Anaconda distribution, available for all major operating systems. You can install anaconda from:

https://www.anaconda.com/download

Anaconda comes prepackaged with many of the core scientific Python libraries, including numpy, scipy, pandas, scikit-learn, and Jupyter. However, there may be other packages we'll cover or use in the course. You should be able to install virtually all of these using the conda or pip installers. From the command prompt, type the following:

conda install [package]

Or (if that doesn't work):

pip install [package]

Packages we'll almost certainly work with (which you may want to install ahead of time) include scikit-learn, seaborn, statsmodels, and rpy2--all of which can be installed with conda install (e.g., conda install seaborn).

Testing your environment

If you want to make sure you have a working scientific Python environment prior to the first day of class (recommended), you can try running the test notebook in this repository. You should be able to view the rendered notebook in your browser even if you don't run it, but I recommend downloading and executing the Jupyter notebook to make sure everything works.

If you're familiar with GitHub, you can clone the entire repository and navigate to the notebook. If you're not familiar with GitHub, the easiest way to execute the notebook is to load the raw notebook code available here in your browser, then save the file to a convenient location (by clicking "Save As..." or the equivalent option in your browser). Now, from the command line, navigate to the directory containing the saved notebook, and run:

jupyter notebook test.ipynb

(If you saved the notebook under a different name, substitute for test.ipynb accordingly.)

After a few seconds, you should see a local version of the notebook open up in your browser. You should now be able to execute the entire notebook by selecting the "Run all" option from the dropdown "Cell" menu. If you run into problems, they'll almost certainly arise in the first code cell containing all of the import statements. The likely culprit is that you've failed to install one or more of the required Python packages, in which case you can go back to the previous section and make sure you've installed Anaconda as well as the other packages mentioned above. If you run into issues (and you're registered in the SSI course!), feel free to email me with questions. If you don't run into any problems, congratulations! You're all set for the course.