Let me start by saying that I did learn some basics of R programming during this course with a lot of help from friends and fellow classmates.
I like to believe that my 20+ professional years in education and training design and delivery have left me with a pretty good understanding of adult learners, learning theory, and putting all of that into practice. With that in mind, here are a couple of points.
First, there is no prerequisite knowledge, skill, or course listed as required for this specialization. Here's what Coursera says about background knowledge, "Some programming experience (in any language) is recommended. We also suggest a working knowledge of mathematics up to algebra (neither calculus or linear algebra are required)." Great! My limited working knowledge of BASIC from the 1980s and novice ability with MS VB fit the bill. Math? No problem, got it covered with Algebra 2 30 years ago. But wait, it turns out neither of those are the case because there are pinned posts in the forum that say the author doesn't understand why an understanding of linear algebra isn't required because it would be really helpful, and you yourself make the point that, "Yes, the mathematics in Statistical Inference and Regression Models are tough for students who haven't previously studied statistics." Knowledge of statistics isn't required or even recommended for this specialization Mr. Greski. That's poor curriculum design and setting students up to fail because there is no realistic expectation set as to what they face in this.
Second, if the materials do not provide any framework or context to tie the assignments to previously taught content either in "lecture", swirl, or assignments, then the course designers and instructors did an incredibly poor job with the design. How can students, even those who have a background in statistics, be reasonably expected to know when an assignment makes use of information or learning that must be found outside of the course itself? This can easily be fixed by including a statement or section with each assignment that says something like, "This assignment covers material found in Lessons, x, y, and z. You will also need information found in sources outside this course such as datacamp.com, etc." The italicized sentence can be the same in every assignment. The paragraph could even say that the assignment is not connected to the current lesson in any way as the intent is for the student to make use of outside resources (or whatever the approriate intent is). Regardless, every graded assignment should have a purpose stating what the student should get out of it, and they could all benefit from a context statement.
Third, if people like me (those with a non-statistics/mathematics background) are not part of the target audience, then please define the target audience better. Currently, Coursera says this, "Beginner Specialization. No prior experience required." That makes it sound like it is appropriate for anyone with no background knowledge or experience because the course will provide all the background knowledge and skills needed along the way. I'm willing to bet that the full program for $3,310 at JHU has prerequisites other than "Beginner Specialization. No prior experience required," and "Some programming experience (in any language) is recommended. We also suggest a working knowledge of mathematics up to algebra (neither calculus or linear algebra are required)."
Finally, portions of this specialization that I have completed so far are out of date. There was one quiz question that involves a specific package in R that is not compatible with the latest version of R. That forces the student to guess. Of course, we can take the quizzes over and over again if we are patient enough so it doesn't really matter if we are guessing on the answers or actually learning and getting the correct answers. Does it? While $49/month may not seem like a lot of money to some people, for others it could be quite a bit if they are having difficulty finding a job and working to improve their skill sets and qualifications. Even though it is "just" $49/month, we are paying for what we presume to be a quality product from a top-notch university. I don't think it's too much to ask that someone correct all the various errors, keep the materials up-to-date with the current version of R (once a year at least), and review feedback such as this (and others) in the forums and Coursera comments. Yes, this means someone has to put the time in on that type of work. I can honestly say that the current state of the materials and quality of design and delivery are well below what I expected for a JHU associated product. I'm not sure I can recommend this to someone as an introduction to data science as it currently exists. I wouldn't be surprised if JHU as an Instructional Design program that could use something like this as a capstone project or similar effort.
I sincerely appreciate the time that the JHU staff, the folks volunteering as mentors, my fellow classmates, and my neighbor give to help me and other students understand the concepts and skills in this course. Our expressed frustration about not seeing a connection between assignments and lectures is not an expressed desire to have our hands held and be spoon fed. It's a frustration at wanting to understand the materials, how they fit together, and how we can use them, which I believe is the intent of education and training in general. Hopefully, Dr. Peng or someone from JHU will see this feedback and be interested in making improvements to the curriculum.