statistics

Jake Vanderplas's PyCon 2016 talk shows how sampling-based methods such as simulation, shuffling and bootstrapping make statistics intuitive. His pitch is that if you can write a for-loop, you can do statistics.

A short Veritasium video explaining regression to the mean and how ignoring it leads to wrong conclusions about punishment versus reward, medical treatments and sports streaks. It uses the example of flight instructors who believed criticism works better than praise.

John Rauser's Strata 2014 keynote argues that programmers can learn statistics quickly by simulating random processes and shuffling data. He explains p-values through resampling rather than formulas.

A four-minute minutephysics video arguing that causal models and causal networks let statistics infer causation from correlation. It touches on Reichenbach's principle, common causes, feedback and quantum causal discovery.

Lecture 1 of Harvard's Statistics 110 with Joe Blitzstein, introducing sample spaces, the naive definition of probability, counting, binomial coefficients and the sampling table.

A two-minute scene from The Wire in which Baltimore police brass pressure subordinates to fix the crime statistics. It dramatizes how metrics and incentives distort policing.

Hans Rosling uses data and animated statistics to show how world population is changing, why birth rates fall as child survival and wealth rise, and why population is expected to peak around 11 billion. It also tackles common misconceptions about Africa and family planning.

Practical Engineering explains what a 100-year flood actually means, combining statistics, hydrology and how society treats risk. Grady walks through local drainage infrastructure and the caveats in estimating flood probability.

Veritasium explains why a large share of published research findings may be false, covering p-values, p-hacking, publication bias and low replication rates. It draws on John Ioannidis's well-known paper and describes reforms such as preregistration.

Hans Rosling and his son Ola run an audience quiz with global population, health and income data to show how wrong most people are about the world. Ola then gives four ways to get less ignorant.

Michael Betancourt's lecture at MLSS Iceland 2014 on efficient Bayesian inference with Hamiltonian Monte Carlo. It covers MCMC concepts and motivation and the problems of sampling in high-dimensional distributions.

A Veritasium video on survivor bias, explaining how an observed sample can be incomplete and skewed. It uses examples such as bombers and famous success stories.

David Kriesel's 36C3 talk, in German, analyzes a year of scraped data on every stop of every long-distance train in Germany. It shows how Deutsche Bahn delay statistics hide cancellations and delays.

minutephysics, in collaboration with the US Census Bureau, explains differential privacy and why the 2020 census adds noise to protect individuals' data. It walks through how statistics can leak private information.

Veritasium explains Bayes' theorem with examples and introduces the 'Bayesian trap', where repeated events get mistaken for inevitable ones. It discusses how priors shape our expectations and behavior.

Matt Parker's LMS popular lecture on how seemingly incredible results can be meaningless random patterns. He ranges from ancient Woolworths civilisations to shoelaces and DNA.

Marcus du Sautoy of the BBC demonstrates the wisdom of the crowd, showing that averaging many guesses of jelly beans in a jar lands near the true count.

SciPy 2015 talk by Nathaniel Smith and Stefan van der Walt on choosing a new default colormap for Matplotlib. It covers why Jet misleads, how perceptual color spaces work, and how viridis was designed.

Joel Grus's JupyterCon talk explaining why he dislikes Jupyter notebooks. He covers hidden state, trouble with reproducibility and testing, and how notebooks frustrate teaching. He shows his preferred workflow and what Jupyter could change.

A five-minute lightning demo of VisiData, a terminal tool for exploring and manipulating data. It uses two months of NYC 311 complaint data to drill down, plot points on a virtual map and save selected rows.

John Kitchin shows how Emacs org-mode with Python lets researchers combine narrative, math, figures and executable code in one document. It exports to LaTeX, HTML and PDF, and he used it for papers and e-books.

Database pioneer Michael Stonebraker argues that 'big data' is really at least four different problems: big volumes with small analytics, big analytics, velocity, and variety. He comments on where each market is heading.

The first video in a 55-part series by Michel van Biezen that introduces the Kalman filter and builds from the rationale through algebraic examples to the full matrix treatment. The videos are short, blackboard-style lectures.

Dan Robinson, then CTO of Heap, explains how Heap runs its product analytics backend on PostgreSQL and CitusDB. It covers conversion funnels, retention and cohort queries across tens of billions of events and 50 TB of data, plus Postgres and Kafka together and the indexing strategy.

A clip from Yes Prime Minister in which Sir Humphrey shows how leading questions can make a survey support or oppose national service. It is a short satirical scene.

David Kriesel's 36C3 talk, in an English translation, analyzes a year of scraped Deutsche Bahn data covering every stop of every long-distance train in Germany. It explores delays and how punctuality is actually measured.

Sandy Pentland of MIT presents his book on social physics, which uses big-data studies of how people interact to explain how ideas spread. He describes findings on collective intelligence and social influence.

Coffee Break analyzes about 40,000 videos from seven months of YouTube's trending tab. It examines whether trending is biased toward mainstream media and late-night shows.

Alexander Nix of Cambridge Analytica presents at the 2016 Concordia Summit on using big data, psychographic profiling and behavioral microtargeting in elections.

Peter Norvig's UBC distinguished lecture on how massive amounts of simple data, such as for language modeling, can yield understanding that hand-built models could not.

Jan Vitek of Northeastern University gives a meetup talk on the challenges of making the R language run fast. He explains what in R's semantics makes it hard to optimize and to build a JIT compiler for.

David Sanders gives the first part of a two-part introduction to the Julia programming language at SciPy 2014. It walks through the language's basics and its appeal for scientific computing.

Bryan Cantrill presents Manta, an object store with in-situ compute that lets tools like grep, awk, and sed run map-reduce style over huge data. He frames it as bringing the Unix philosophy to big data.

Prof. Eran Segal's TEDx talk on research showing that people's blood sugar responses to identical foods vary widely. It describes an algorithm that predicts individual responses and why one best diet for everyone does not exist.

Excerpts from W. Edwards Deming's Red Bead Experiment, where workers draw beads from a bowl and are judged on red ones that are largely down to chance. It shows how system variation, not individual effort, drives results.

Freeman Dyson recounts how Enrico Fermi rejected his team's meson scattering model, noting that it had too many free parameters. It is the source of the quote about fitting an elephant.

A scene from Moneyball in which Brad Pitt's character argues with scouts who pick players on intangible impressions instead of performance data. It is a four-minute clip.

Clip from a John Oliver segment showing Newt Gingrich arguing that feelings about crime matter more than FBI statistics at the 2016 RNC.

CGP Grey explains first-past-the-post voting and why it pushes elections toward two dominant parties. It is the first in his series on electoral systems.

A short CGP Grey explainer on the alternative vote (instant runoff voting) versus first past the post, made around the 2011 UK referendum. It shows how ranked ballots are tallied and why they avoid spoiler effects.