edX: HarvardX Data Science: Linear Regression Review

Data Science: Linear Regression is the 7th course in the 9-part data science professional certificate track offered by Harvard on the edX course platform. The course covers the basics of topics related to correlation and linear regression with a focus on creating models in the R programming language. Although this course is the 7th part of a series, you don't really need to have taken the previous courses to take this one: all you need is basic knowledge of R. This course dispenses with the programming assignments administered through DataCamp featured in the first four courses of the specialization and instead grades students via multiple choice questions after the lecture videos.

Linear Regression divides course content across three main topic sections: Introduction to Regression, Linear Models and Confounding. The first discusses the concept of correlation and stratification, while introducing some terminology and concepts from the sport of baseball as a motivating example for how using data-driven methods have helped make decisions in the past. The second section covers the basics of linear regression and how to create a regression model in R as well as some considerations related to R data structures. The final section touches on issues that can confound models and lead to inaccurate conjectures when working with linear models and correlation. 

The video and instruction quality in the course are very good and the videos cover the topics in about the level of sophistication that would be expected given the expectations set by the previous courses in the specialization. You'll learn how linear regressions models can be used and how to create them with R, but as a professional certificate track, don't expect to get too deep into details. The after-lecture multiple choice questions help reinforce the video content, but it would be nice if there was at least a couple of programming assignments to allow students to really dig into the material with R. As it stands now, you can complete the entire course without actually using R yourself as long as you know it well enough to answer the multiple choice questions.

Data Science: Linear Regression is a good basic introduction to linear regression that students who have been following along with the data science certificate track should be more than ready to tackle. It is too bad that the course creators decided to stop offering programming assignments, as the DataCamp projects were the highlight of the first four courses in this series, but with a topic like linear regression, conceptual understanding is more important than programming, so the lack of coding is more excusable than it was in the previous course on data wrangling.

I give Data Science: Linear Regression 3.75 out of 5 stars: Good.