Correlations, Associations and Hypothesis Testing (with R)

Name: Correlations, Associations and Hypothesis Testing (with R)
Rating: 4.6 (96 reviews)

Data Science Prerequisites: Correlations, Associations and Hypothesis Testing (with applications using R)

Highest Rated

Created byViani D.B.

Last updated 5/2022

English

English [Auto],

What you'll learn

Exploratory analysis for association between variables/features.
Foundation and essence of hypothesis testing for association between variables/features.
Measure and quantify the strength and statistical significance of associations.
Applications using real world datasets.

Course content

5 sections • 28 lectures • 2h 53m total length

Motivation and objectives3:39
Discover how correlations and associations drive data-driven decisions using R, with practical sessions on numerical variables, categorical variables, and numerical and categorical analyses, including hypothesis testing and interpreting p-values.
Extract from detailed objectives (with illustrations)5:15
Getting started with R/RStudio (Optional)6:47

The covariance: Interpretation and limitations10:25
Explore how covariance measures whether two variables move together, how to compute and interpret sample covariance, and the limitations of scale sensitivity that lead to using the correlation coefficient.
Correlations2:42
Explore how the population correlation coefficient standardizes covariance by the product of standard deviations, revealing the magnitude of the linear association between X and Y.
Illustration and interpretation6:38
Watch how the correlation coefficient captures the strength and direction of the x–y association as data patterns move toward a straight line and toward perfect positive or negative correlation.
Estimation14:08
Compute the correlation between X and Y by estimating sample variance and covariance, then derive the correlation coefficient in R via manual steps or the cor function.
Lab 1 (Task specification and datasets)1:15
Solution to Lab 119:48
Hypothesis testing9:51
Assess whether a sample correlation indicates a true linear association by testing H0: rho = 0, versus H1: rho ≠ 0, using the t statistic and p-values with alpha.
Performing the test in R and interpretation9:51
Lab 2 (Task specification and data file)0:27
Solution to Lab 27:54
Association between numerical variables

Introduction3:13
Explore how to assess association between categorical variables using a cross-frequency table in R, illustrated by income class and gender, including sample size, percentages, and significance questions.
Special case of categorical variables with two categories11:11
Assess whether income class and gender are independent using the chi-square test, comparing observed and expected cross-frequency tables under H0 to detect association.
General case and hypothesis testing6:40
Evaluate independence between two variables by constructing cross frequency tables, computing expected frequencies, and using the chi-square test to derive a p-value for deciding H0 vs H1.
Performing the test in R and interpretation1:36
Lab 3 (Task specification and data file)0:27
Solution to Lab 39:38
Association between categorical variables

Introduction1:13
Learn how to assess the association between a categorical variable and a numerical variable, using examples like gender and BMI, and apply these methods in R.
Hypothesis testing5:01
Compare mean medical charges for males and females to test for association with a t-test. Compute the statistic from sample means and s, and reject H0 if p < alpha.
Assumptions2:04
Identify the conditions for testing association between numerical and categorical variables, including normality via the Shapiro test and homogeneity of variances via the Bartlett’s test.
Performing the test in R and interpretation3:03
Perform a t-test in R to compare mean medical charges for males and females, reporting t, df, and p-value. With p about 0.61, we fail to reject H0—no significant difference.
ANOVA and the F-test3:41
Learn how anova extends association testing from two-category to multi-category categorical variables, using the f-test to compare means across q categories of a numerical variable, and interpret p-values.
Lab 4 (Task specification and data file)0:22
Solution to Lab 422:36
Association between numerical and categorical variables

Requirements

No prior experience is required. We will start from the basics.

Description

Exploring and assessing the strength of associations between variables/features plays a fundamental role in statistical analysis and machine learning.

I decided to create this course after leading many data science projects and coming across many data scientists struggling with the fundamentals of association between variables/features and hypothesis testing.

This course will be beneficial to junior analysts as well as to more experienced data scientists. In particular,

If you are an aspiring/junior data analyst/scientist, this course will contribute towards building the right foundation at an early stage of your career.

If you are an experienced data scientist, this course will help you to re-visit and eventually improve your understanding of the assessment of associations between variables/features.

The course is divided into three main sections.

The first section looks at the assessment and quantification of associations between numerical variables.

The second section focusses on the assessment of associations between categorical variables.

The third section covers the assessment of associations between numerical and categorical variables.

Each section discusses a number of statistical metrics in relation to associations between variables and then build statistical hypothesis tests to measure the strengths of these associations.

There are practical sessions throughout the course, where you will see how to implement the methods discussed in the course (using R) and to perform various hypothesis testing using real world datasets. Your will also learn and master how to interpret results in a broader context.

In addition, quiz is added at the end of each section. The objective of these quizzes is to help you to consolidate the main concepts covered in the course.

By the end of the course, you will have a clear and coherent understanding of covariances, correlations, t-test, Chi-squared test, ANOVA, F-test, and much more. In particular, you will know when to use these tests and how to ensure that the underlying assumptions are satisfied.

Who this course is for:

Anyone interested in data analysis.
All levels

Correlations, Associations and Hypothesis Testing (with R)

What you'll learn

Explore related topics

Course content

Motivation and objectives3 lectures • 16min

Association between numerical variables10 lectures • 1hr 23min

Association between categorical variables6 lectures • 33min

Association categorical and numerical variables7 lectures • 38min

Concluding remarks2 lectures • 4min

Requirements

Description

Who this course is for: