the materials are out dated like hell, but meh at least i got the basics

Really was some painful experience. I did manage to get over the first two weeks battling myself in those overrated assignments. But after the 3rd week it was really like a typical lecture in the university where professor reads out his slides and give the assignment to check whether the student was able to stay awake the whole time of his presentation. It was really hard to catch up the gap between those couple of weeks. Not recommended for a beginner or a person who has just entered the field of DS. Believe me, you're in for a really grumpy ride with this course.

The course could be much better if somebody would read through the discussion comments and make appropriate changes to the wording in lab assignments.

I would not have completed the lab assignments without reading through several of the course discussion comments. Many of these comments have been there for years, yet no improvement to the course material.

Very theoretical, with good slides but missing good Jupyter notebooks that facilitate the "hand-on" experience. Thus, it's up to the students to find out their own way to solve the assignment. Courses 1, 2, 3 of the Applied Data Science with Python Specialization are much, much better than what is offered in Course 4.

Week 2 and Week 3 are OK. But Week 1 and Week 4 are Horrible. All the other courses in this specialization are amazing. But this one, I don't think so. If you really want to learn NLP, go for other sources. This course is just like a revision for the ones who are already pretty good with NLP.

The lectures are OK, but the assignments have very bad instructions and many bugs. Expect to spend 5% of the time learning and 95% trying to figure out what the autograder expects as an output. Overall there are much more efficient ways to learn. I can't recommend this course.

The auto-graders are not working for quite a few questions, and error description does not tell which part of my code fails, totally useless description. Discussion forum was a big disappointment. Mentors never showed up. I felt like this is a free course, in terms of support.

I have taken 4 courses thus far in this specialization. Course 4 is the most horrendous.

If you embrace sadistic tendencies, please take this course.

If you like to be taken for a ride and waste valuable time, please take this course.

If you like to be trolled by the instructor, its so called course syllabus and the champion of them all, the beloved autograder time and time again, please take this course.

I will specifically mention Assignment 4. The course syllabus does not even cover fully the material that will be tested in the final assignment. Its a lot of searching stack overflow and reading documentation to implement the answer. That is not all. The functions and the way it is laid out to be accepted as an answer by the autograder is ridiculous. When there are errors, the autograder doesn't even suggest what the error is. It just tells you the answer is incorrect. How does one debug? I mean.. wtf right?!

Nonetheless, I will push on and finish this specialization because it is a commitment to finish what I started.

All in all, course 4 is unforgivable. This is not the way to teach and not the way for learners to learn.

Broken. That's the only word I can think of. The course ...and me in the process.

I found this course quite confusing and often unrelated between video lectures and assignments. The lectures maybe covered an assignment in broad strokes but to actually answer any of the questions needed extension research from the student. I felt like I was teaching myself the base content. At that point, what is the point of the lecture videos if they provide no value. I almost stopped my subscription and gave up on the data analysis specialization based on the quality of this specific course. Previous courses in the specialisation did provide useful information in lectures which was then extended upon in the assignments. This method of teaching something in the lectures then building on finessed usage in the assignments is a much better approached.

The lectures for this course are terribly uninspired, giving very little useful information--the vast majority of it is the professor talking about obvious aspects of language at a very high and useless level. The autograder is frequently breaking for very minor things (such as returning numpy.float instead of float), the questions on the assignments are often misleading, poorly worded, vague, or just generally not very helpful. All in all, this was one of the worst MOOCs I have ever taken, though the Coursera bar is pretty low. It does make me wonder why I bother to pay at all--oh right, Coursera now makes not paying a major inconvenience to course progression.

This is my least favourite course in the specialisation. Natural language processing is an exciting field and I think there is a lot more potential to enthuse and engage students. The instructor scratches the surface of text mining by going through brief sets of codes on ppt slides. I thought it would be meaningful to use more real-world datasets (as in the previous courses in the specialisation) and have students follow through some examples on Jupyter Notebook. I also felt that the exposition by the instructor was not the most intuitive or lucid. It could be much clearer.

While containing marginally less bugs and errors than previous courses in the specialization (though they are still present and bafflingly U of M shows no interest in ever fixing them), this course spends too much time going over topics already covered in previous courses, then shifts to a very simplistic and rudimentary overview of the material at hand. The last week especially is so rushed that I don't feel I have a solid grasp on what was supposed to be taught, and the assignment is just paint-by-numbers reading of documentation and plugging in guesses.

This course was much less helpful than others in the Specialization. The assignments are poorly conceived, and submissions are beset by finicky autograder issues. Certainly, data cleaning and code debugging are critical skills for text mining, but I find it difficult to believe that "try to understand what output a function should submit so as to satisfy the current autograder" is a useful way to teach text mining.

I hope this course will be re-done to bring it in line with the quality of the others in the Specialization.

The course is a joke. Its outdated and not supported, you literally need to spend hours to try and figure and emulate versions used by autograder and even the file structure for files used by default is not accurate and you get file read errors on predefined by them functions on their own virtual environment and need to fix these for them!!! The virtual machine env provided is super slow so need to use your own. Very bad user experience and horrible use of time!

What a horrible course. Especially the assignments are such an unbelievable waste of time. Instead of focusing on important concepts and applications, one has to spend hours one "pleasing the autograder" by renaming columns and reading the discussion pages for the correct interpretation of all the ambiguously formulated questions. Very sad! Would be good for everyone if this was removed from the (otherwise great) series "Applied Data Science in Python".

I was under the impression that the course is incomplete, especially week 4, which has no notebook examples of the theory presented. I needed to look at other sites for basic information. I could only complete the exercises because they are easy, otherwise, with the code presented during the course, I would not have been able to. I suggest strengthening the example code in python (see week 3, good code)

Do yourself a favor and do the NLP specialization from deeplearning.ai. This course (and specialization in general) is outdated. Assignments are poorly organized and the poor moderators can only warn you about assignments as they're unable to change anything. After taking Dr. Chuck's & Dr. Resnick I was seriously dishes out some cash on the online Master's they're offering. Glad I didn't...

I have to say that the previous three courses were very well explained, with good examples and python code. However, this course is not well explained nor documented. It is a pity that the quality of the whole specialization program gets considerably reduced due to this course. The assignments do not allow you to learn and you may not pass them due to small differences in the coding.

Videos are so high-level that they don't help at all understanding the necessary code. Assignments have spelling errors and ambiguity. Week 4 is missing the sample code notebook. I eventually found the sample code notebook in the forums, but this was a big cause of frustrations as I had zero context for how to do the assignment.

I don't feel like I learned anything even though I passed with 100%. This course desperately needs more insightful quizzes and assignments, and the lectures should actually explain how to do "applied text mining" rather than just glossing over some terminology.

Unless the instructional staff updates the programming assignments to reflect updates in packages and ensures they can run without additions, do not take this course. It is a terrible reflection on the University of Michigan.

The subject is interesting however the lectures are too shallow and the assignments too difficult. You should be expected to do more study after the lecture for sure but for this course, it was a lot.

Not a great course. I'd skip it. The assignments were just trying out different parameters. Nothing related to machine learning/using Python was discussed in the class (may be 2%). Didn't lean much.

Video lectures are just been run through. No clear explanation at all. On the top of that, assignments are freaking difficult being totally irrelevant to the material taught in the video lectures.

Interesting topic, but a really poor course with barely any content.

Around an hour or less of lectures a week.

I've taken a lot of MOOC's on Coursera and other platforms and this one is poor