Automating Analysis
summary
Scrapes of HowSTAT and NCERT surveys drive concrete claims: opponent country explains 16% variance in cricket strike rates, pitch and home country follow, and day-of-week effects are statistically robust at p < 1e-8. Education data uses ~200,000 students and ~1,000 questions in spreadsheets and Python; calculators associate with about calculators +3pp, computers -8pp in math, while parental education is the biggest driver. Methods demonstrated: pivot tables, k‑means, decision trees, ARIMA, ensembles and neural nets, D3/SVG visualizations; automation surfaces results, humans handle interpretation.
AI-written summary · about
chapters (19)
- 0:00Intro
- 1:40How do you get better
- 2:40Variation
- 3:38Ground
- 8:57Data
- 10:52Correlation
- 12:08Computers
- 16:07Statistical significance
- 18:38What makes a difference
- 20:43Algorithmic approach
- 22:58Analyst role
- 26:09Interpretation
- 29:05Identifying clusters
- 31:49How innovation is clustered
- 33:25Where is Saudi Arabia
- 35:28Analyst vs Data Scientist
- 37:21Technology stack
- 38:03Visual analytics
- 40:32Sentiment analysis
related talks
48mHuman data interactions: experiments in how humans analyze data
39mSoftware Analytics with Data Science on Software Data
01h 53mAnalyze your data in Python
02h 53mMining digital exhaust
32mLearning Statistics Will Save Your Life
44mPython for Economists: An overview of Python tools for Economists
35mExposing Algorithms
41mFootball's data analytics in python