Survival Analysis in Python and R
summary
Uses Python Lifelines and R libraries survival and KMsurv, pulls author timelines from arXiv and enriches with genderize.io and affiliation parsing, fits Kaplan‑Meier curves and Cox models, and compares random survival forests and GAMs. Dataset narrows to ~4,244 authors, median academic lifetime ≈ nine years, concordance around 0.60–0.69. Code extracts durations, handles censoring, reports which covariates (papers in PhD, coauthors, institution) move hazard rates and supports Lyst-style out‑of‑stock prediction.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 3:16Survival analysis pipelines
- 7:12Survival analysis basics
- 10:44Academic career data
- 17:27Cleaning survival data
- 23:09Nonparametric estimation in Python
- 26:59Model performance with C-index
- 29:23Big data with GAMs
- 30:27Random survival forests
- 31:50Key findings
- 32:31Q&A
related talks
32mReproducible Research in Python
01h 24mIntroductory tutorial on data exploration and statistical models
45mPython useRs
27mTweaking the rise and fall of empires and economies
36mImplementing and Training Predictive Customer Lifetime Value Models in Python
05mA crash-update to lifelines
33mPython and Data Storytelling to the Rescue: Let's Avoid More Deaths by...
18mTo Explore or To Exploit: Decoding Human Decision Making with R and Python