STAT1021: Introduction to Data Science (Spring 2018)

Announcement (updated Apr. 21 16:30)

• The networkie library is updated (Please do git pull origin master often). It now includes the raw result of the friendship questionnaire and the script that parses the data into a neat csv file.
• There's a crowdsourced list of awesome GitHub projects for the lecture. Check it out here! And, help improve the data by adding new repos that you like.
• The slides on Apr. 18 were adapted from Prof. Aaron Clauset's slides for the 2015 SFI Networks short course.
• The slides on Apr. 11 is here.
• Please install a shell emulator (if you are using Windows) and Git on your computer. You may follow this link for instructions.
• Read 鳥哥的 Linux 私房菜 if you are unfamiliar with shell commands.
• Please sign up a new GitHub account on GitHub. You might clone or fork a project that you are interested.

Lecture Description

This short lecture series provides an introduction to the daily practices and workflow for a data scientist/engineer in the industry, with an emphasis on building intuition on network data. Specifically, we will sample and analyze the in-class social network and introduce broadly applicable concepts in network analysis.

The lectures will be delivered as three integrated parts: on using Git and GitHub, on website hosting on GitHub Pages, and on a very short (but not hurried) introduction to Network Science. These will include (web) software development methods and tools; graph theory and graph concepts; community detection; the implications of popular concepts such as "six degrees of separation" and the "friendship paradox".

After these lectures, the students will be able to publish an open source software on GitHub that analyzes and visualizes network data and host a personal/professional website on GitHub Pages.

Important links:

• STAT1021 official webpage
• Detailed outline
• Python library "networkie" on GitHub

Prerequisites

The materials will be accessible to students with little programming experience; however, they must be able to write and run Python scripts on the command-line. This tutorial might also appeal to people being interested in Network Science. Yet it will focus on the practical side, i.e. using Python for network data analysis. Web development experience is not required but a prior exposure to HTML/CSS/JavaScript would be useful.

Problem Sets

Problem sets are posted on the course website (here!), solutions will be sent through email soon after the submission deadline. See the detailed outline for information about grading, late, and collaboration policy.

Data

In-class social network, collected via the self-reported friendship questionnaire (in Mandarin Chinese).
Zachary Karate Club

Tentative Schedule