John Perry Barlow on politics and the internet (1998) minireference 2021-12-29 | This is an excerpt from 16:14 until 17:44 from the video C-SPAN source video c-span.org/video/?110679-1/politics-government-digital-technology
Sec 3.2.3 Confidence intervals for the population mean minireference 2023-08-16 | In this video we'll construct confidence intervals for the mean of an unknown population based on a sample taken from that population.Notebook: nobsstats.com/notebooks/32_confidence_intervals.html#confidence-interval-for-the-population-mean
Sec 3.2.5 Confidence intervals for the difference between means minireference 2023-08-16 | We show how to compute confidence intervals for the difference between means of two unknown populations. We'll use an analytical approximation based on Student's t-distribution and the bootstrap approach.Notebook: nobsstats.com/notebooks/32_confidence_intervals.html#confidence-interval-for-the-difference-between-means
Sec 3.2 Confidence intervals minireference 2023-08-16 | Introduction to confidence intervals. Every self-respecting statistician calculates the uncertainty of the estimates they compute.Notebook: nobsstats.com/notebooks/32_confidence_intervals.html
Sec 3.2.4 Confidence intervals for the population variance minireference 2023-08-16 | We construct confidence intervals for the population variance based on one sample of observations. We'll show constructions both on the chi-square distribution and bootstrap estimation.Notebook: nobsstats.com/notebooks/32_confidence_intervals.html#confidence-interval-for-the-population-variance
Chapter 3 Inferential statistics intro minireference 2023-03-29 | Chapter 3 is all about inferential statistics. Suppose you have collected the sample (x1,x2,...,xn) from a population. You assume the probability model X ~ ℳ(θ) describes the population, where ℳ describes the model family (e.g. uniform, Poisson, normal, exponential, etc.) and θ are the unknown model parameters that characterize the specific population you're interested in. Statistical inference is a bunch of tricks that allow you to use the observed data in the sample (x1,x2,...,xn) to compute an estimate θ̂ that approximates the unknown parameter θ.See minireference.com/static/excerpts/noBSstats/conceptmaps/BookSubjectsOverview.pdf for the concept map, and nobsstats.com/notebooks/31_estimators.html for the first notebook.
Sec 3.1 Estimators minireference 2023-03-29 | Estimators (functions computed from samples) are the essential building blocks you need to know to understand statistics. This is why we'll start with this topic, which is the most complicated, and cognitively demanding in the book.The good news is, once you understand estimators, you'll know everything you need to know to understand statistics in a deep way. Once you learn the "math machinery" of estimators and their sampling distributions, all the rest of the chapter (confidence intervals and hypothesis testing) will be smooth sailing.See nobsstats.com/notebooks/31_estimators.html
Sec 3.1.3 Sampling distributions minireference 2023-03-29 | The sampling distribution of the estimator g, is the distribution of estimates we might observe when we compute the estimator on random samples. This is the most complicated concept in this chapter, so in this video you'll see me try my best to use visuals, math, and code to explain it.See nobsstats.com/notebooks/31_estimators.html#sampling-distributions
Sec 3.1.4 Approximating sampling distributions minireference 2023-03-29 | In realistic statistical inference scenarios, the population parameters are unknown, so if we want to report the uncertainty in our estimates, we must learn find some way to approximate the sampling distributions from observed data (e.g. from a single sample).This lesson introduces the bootstrap estimation procedure (a computational approach) and the analytical approximation methods (based on math formulas) for approximating sampling distributions.nobsstats.com/notebooks/31_estimators.html#approximating-sampling-distribution
Sec 3.1.5 Sampling distribution of the mean minireference 2023-03-29 | This lesson is a step-by-step introduction to the various methods for estimating the sampling distribution of the man. In particular, we'll show an analytical approximation based on Student's t-distribution (very good for small samples coming from a normally distributed population). We'll also show how to obtain an approximation to the sampling bootstrap of the mean using bootstrap estimation.See nobsstats.com/notebooks/31_estimators.html#id2
Sec 3.1.6 More sampling distributions: var, dmeans, Explanations and Discussion minireference 2023-03-29 | Quick overview of the analytical approximations and bootstrap estimates for the sample variance and the difference between means. This video also includes a walkthrough of the Explanations and Discussion subsections.See nobsstats.com/notebooks/31_estimators.html#id3 Exercises notebook mybinder.org/v2/gh/minireference/noBSstatsnotebooks/HEAD?labpath=exercises%2Fexercises_31_estimtors.ipynb Try them! Remember, statistics is not a spectator sport!
04 LINEAR MODELS minireference 2021-11-22 | Linear models allow us to describe relationships between two continuous quantities x and y. We assume the output Y is distributed according to Y ~ b+m*x, in other words the x-Y relationship is a line of the form y=mx+b. In this video we study the case when each employee receives a variable amount of stats training (anywhere from 0 hours to 100 hours), and the ELV value is modelled by the relationship ELV ∼ 𝛽0+𝛽1*hours, where 𝛽0 is the intercept and 𝛽1 is the slope. The notebook shows several different ways to fit the model (find the best choice of parameters 𝛽0 and 𝛽1), interpret the results, and make predictions.Notebook: mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F04_LINEAR_MODELS.ipynb
03 STATS (part 5/5) - Confidence intervals for the effect size minireference 2021-11-22 | In this video we'll learn how to provide rigorous estimates of gains in ELV for employees that complete the stats training, which is called effect size in statistics. We'll learn two different ways of estimating the effect size: one based on resampling methods (bootstrap) and another based on analytical approximations and Student's t-distribution. Both approaches give us a confidence interval, which describes the range of values for an effect size we can expect, and we can give a 90% guarantee on the procedure we use to generate the interval.Notebook: mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F03_STATS.ipynb starting at "Estimating the effect size"
03 STATS (part 4/5) - Hypothesis testing using the t-test minireference 2021-11-22 | Analytical approximation methods in statistics provide ready-made formulas for describing the sampling distribution in various statistical analysis scenarios. These formulas are normally taught in STATS 101 courses because they require very few steps and can be applied without needing to understand the math models. For the specific scenario of comparison of two group means, we can use Student's t-test (so called because it is based on Student's t-distribution). This video shows how to apply Student's t test to obtain the p-value 0.047 and reject the null hypothesis.Notebook: mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F03_STATS.ipynb starting at "Approach 2: Analytical approximations for hypothesis testing"
03 STATS (part 3/5) - Hypothesis testing using the permutation test minireference 2021-11-22 | The permutation test allows us to obtain the sampling distribution under the null hypothesis H0 by using existing sample data and a clever "shuffling" of the group labels. Using this trick, we obtain the sampling distribution of the difference between group means according to the probability model for H0. We then find where the observed difference between group means d=130 falls within the sampling distribution to obtain the p-value. Since the p-value we obtain is less than 0.05 our conclusion is to reject the null hypothesis, i.e., we conclude that there is a statistically significant difference between the groups means.Notebook: mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F03_STATS.ipynb starting at "Approach 1: Permutation test for hypothesis testing"
03 STATS (part 2/5) - Hypothesis testing intro minireference 2021-11-22 | Hypothesis testing is an approach to formulating research questions in the form of yes-or-no decisions. Hypothesis testing is a standardized procedure for doing statistical analysis, and a common tool use din academic research papers. We formulate research question as two competing hypotheses: - H0, the null hypothesis, which is a baseline, skeptical claim like "no difference" or "zero effect", and- HA, the alternative hypothesis, which usually embodies a theory or model that the researchers want to study.In order to justify any study or analysis of the alternative hypothesis HA, scientific convention requires us to first "reject the null hypothesis," which we can do if we can show that the observed data is very unlikely (less than 5% chance) to have occurred by chance under H0. The logic is if the data is unlikely to occur under H0, this lends support to HA and further study is warranted. The alternative is that we "fail to reject the null hypothesis," which means the data observed is consistent with the "no difference" or "zero effect" data model, so does not require postulating any alternative hypothesis.Notebook: mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F03_STATS.ipynb starting at "Hypothesis testing"
03 STATS (part 1/5) - Intro and Estimators minireference 2021-11-22 | Estimators (functions 𝑓 computed from samples) are a fundamental tool for statistics. The descriptive statistics we saw in the DATA chapter are examples of estimators: the sample mean is an estimator for the population mean, and the sample variance is an estimator for the population variance. We denote by 𝑓(𝐱) value of the estimator computed from a particular sample 𝐱. We denote 𝑓(𝐗) the value of the estimator computed from a random sample 𝐗, and refer to the probability density function of 𝑓(𝐗) as the sampling distribution of the estimator 𝑓.This video provides a detailed hands-on discussion about the difference-between-group-means estimator `d = mean(xA) - mean(xB)`, which will is the math building block we'll need for both hypothesis testing and estimation. We'll obtain the sampling distribution for the difference-between-group-means estimator using both simulation and mathematical modelling techniques.Notebook: mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F03_STATS.ipynb
02 PROB - Probability theory minireference 2021-11-22 | This video introduces some fundamental topics from probability theory like random events, random variables, and probabilistic models. Understanding probability theory allows us to build mathematical models for the variability in data we observe, and we're lucky to have the lego pieces in `scipy.stats.distributions` docs.scipy.org/doc/scipy/reference/stats.html#probability-distributions at our disposal. Using the code from `scipy.stats.distributions` we can model most common random variables, generate random samples, do calculations using the probability functions and their inverses, etc.The second half of the video is concerned with that probability distribution for samples of size n from a probability distribution. We study the properties of certain statistics (quantities computed from a sample) and learn about the Central Limit Theorem which tells us some useful facts about the distribution of the sample mean (usually denoted as overline-x) from any distribution.Notebook: mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F02_PROB.ipynb
Introduction and 01 DATA minireference 2021-11-21 | In this video, we'll introduce the running example we'll use for the whole video series: Amy's experiment about providing additional training to her startup employees. In the second part of the video, we'll illustrate concepts from chapter 01 DATA like extracting data from a spreadsheet, transforming data from "wide" to "tall" format, and viewing descriptive statistics like histograms, box plots, strip plots, and numerical summaries.Notebooks:- Intro mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F00_Introduction.ipynb - 01 DATA: mybinder.org/v2/gh/minireference/noBSstatsnotebooks/main?labpath=stats_overview%2F01_DATA.ipynb
Linear algebra overview minireference 2021-07-01 | An overview of the material from the *No bullshit guide to linear algebra* http://gum.co/noBSLALinear algebra is the study of vectors and linear transformations. This video serves to introduce the computational, geometrical, and theoretical aspects of linear algebra.You can view the notebook used in this video here:http://nbviewer.jupyter.org/github/minireference/noBSLAnotebooks/blob/master/Linear_algebra_chapters_overview.ipynb
Chapter 2: Linearity intuition minireference 2021-07-01 | This video introduces the main idea of linear algebra: linearity. A function f is linear if it obeys f(ax+by)=af(x)+bf(y) for all inputs x and y, and all coefficients a and b.Linear algebra is the study of "linear structure" in all its forms and shapes. This video explains the concept from the ground up using a few examples, and shows the connection between linear transformations and the matrix-vector product.Notebook: nbviewer.jupyter.org/github/minireference/noBSLAnotebooks/blob/master/chapter02_linearity_intuition.ipynb
Chapter 2: Vectors and matrices definitions minireference 2021-07-01 | Definitions of vectors and matrices—the fundamental objects of linear algebra. This video defines the basic operations we can perform on vectors and matrices: addition, subtraction, products, etc.The Jupyter Notebook for this video is http://nbviewer.jupyter.org/github/minireference/noBSLAnotebooks/blob/master/chapter03_definitions.ipynb
Chapter 3: RREF theory and exercises minireference 2021-07-01 | This video introduces the Gauss-Jordan elimination procedure for solving systems of linear equations through a series of examples. The solutions are illustrated graphically and the special cases of no-solutions and infinitely many solutions are discussed.Notebook is here: http://nbviewer.jupyter.org/github/minireference/noBSLAnotebooks/blob/master/chapter03_exercises.ipynb
1995 Technology In Education hearing featuring Seymour Papert and Alan Kay hearing on education minireference 2020-12-28 | A 1995 heating by House Science & Educational Opportunities committees in which Seymour Papert and Alan Kay give interesting opinions on how technology can be used to improve education.
Writing a book in two minutes minireference 2017-03-02 | Writing a textbook takes a lot of time, but it's totally worth it if you think you have something to teach. This book shows three years worth of writing and edits on the No bullshit guide to linear algebra.The video was recorded using OBS and the visualization was generated using `gource -s 0.05`, which means 1 day = 0.05 seconds of video.minireference.com
