Tabs, spaces and salaries: a data science detective story
summary
Analysis of the Stack Overflow developer survey shows developers who report using spaces have about a $10,000 difference in median salary across experience levels, but a linear regression with country, experience, language, education, open source and VCS yields adjusted R^2 ≈ 39% and removing indentation barely changes fit, so indentation is a proxy not a cause. Investigation finds strong links to contributing to open source and using git vs TFS, plus widespread salary reporting errors (~25% in some countries); data and code are on GitHub.
AI-written summary · about
chapters (7)
- 0:00Start
- 2:35CORRELATION AND CAUSATION
- 15:39COLLINEARITY?
- 18:29POTENTIAL EXPLANATION?
- 20:26LANGUAGE AND OPEN SOURCE?
- 27:34WHY IS VERSION CONTROL SO IMPORTANT?
- 43:21BEYOND "CORRELATION IS NOT CAUSATION"
related talks
45mTabs, spaces and salaries: a data science detective story
52mTabs, spaces and salaries: a data science detective story
01h 04mWhat can we learn from 750 billion GitHub events and 42 TB of code
55mExploring StackOverflow data
01h 03mWhat can we learn from 750 billion GitHub events and 42 TB of code
35mRegulations, Ethics And Messy Data: Tales From An Open Source Analyst
18mData Analysis with JavaScript
25mScope of data journalism for developers