Fitting structure to data without fooling yourself.
10 topics · 14 curated works
No prior grounding assumed.
Understanding Survival Analysis: Kaplan-Meier Estimate
Manish Kumar Goel, Pardeep Khanna & Jugal Kishore · 2010
Walks a non-statistician reader through reading a Kaplan-Meier curve and the log-rank test used to compare two of them.
+3 more at this level
Assumes you know the vocabulary.
Regression Models for Count Data in R
Achim Zeileis, Christian Kleiber & Simon Jackman · 2008
Shows that standard Poisson and negative-binomial regression systematically mis-fit real count data, and works through the hurdle and zero-inflated…
+4 more at this level
Primary sources and full treatments.
Nonparametric Estimation from Incomplete Observations
Edward L. Kaplan & Paul Meier · 1958
Derives the product-limit estimator, showing how to estimate a survival curve correctly when some subjects are censored rather than fully observed.
+4 more at this level
12 of 14 works
Manish Kumar Goel, Pardeep Khanna & Jugal Kishore
Walks a non-statistician reader through reading a Kaplan-Meier curve and the log-rank test used to compare two of them.
link checked 17 Sept 2026Rob J. Hyndman & George Athanasopoulos
Teaches forecasting as a practical workflow — decompose, model the remainder, validate out of sample — with the code needed to carry out each step.
link checked 17 Sept 2026Xavier A. Harrison, Lynda Donaldson, Maria Eugenia Correa-Cano, Julian Evans, David N. Fisher, Cecily E. D. Goodwin, Beth S. Robinson, David J. Hodgson & Richard Inger
A practitioner's walk-through of when a mixed-effects model is actually warranted, and the misspecifications, pseudoreplication chief among them, that make one necessary.
link checked 17 Sept 2026Andrew Gelman, Jennifer Hill & Aki Vehtari
Teaches regression as an iterative loop of building, fitting, checking and revising a model against messy real data, rather than a single closed-form calculation.
link checked 17 Sept 2026Achim Zeileis, Christian Kleiber & Simon Jackman
Shows that standard Poisson and negative-binomial regression systematically mis-fit real count data, and works through the hurdle and zero-inflated models built to fix it.
link checked 17 Sept 2026Sébastien Lê, Julie Josse & François Husson
Argues that PCA, correspondence analysis and clustering belong in one package because they answer the same question, how to summarise a table of measurements, for continuous, categorical and mixed data respectively.
link checked 17 Sept 2026Rob J. Hyndman & Yeasmin Khandakar
Argues that automatic model selection for ARIMA and exponential smoothing, done properly by search over a stated criterion, beats an analyst eyeballing an ACF plot for routine forecasting at scale.
link checked 17 Sept 2026Gareth James, Daniela Witten, Trevor Hastie & Robert Tibshirani
The approachable companion to Elements: the same models with the proofs traded for intuition and worked examples.
link checked 17 Sept 2026Edzer Pebesma & Roger Bivand
Argues spatial data need their own representations and models because location breaks the independence assumption ordinary statistics depends on.
link checked 17 Sept 2026Edward L. Kaplan & Paul Meier
Derives the product-limit estimator, showing how to estimate a survival curve correctly when some subjects are censored rather than fully observed.
link checked 17 Sept 2026David R. Cox
Introduces the proportional hazards model, letting covariates shift a survival curve without ever having to specify the baseline hazard itself.
link checked 17 Sept 2026Robert Tibshirani
Proposes penalising regression coefficients by their absolute value rather than their square, which shrinks some coefficients to exactly zero and performs variable selection as a side effect of fitting.
link checked 17 Sept 2026