I am surprised that it was a 4-week course. It could have been condensed to 1 slide stating "Go read a few books and if you're stuck, consult Stackoverflow".

I cannot believe the average score that this course got and I mean that literally: I can't believe those are real numbers.

There are no course materials, except some crappy unformatted transcripts, the speed is ridiculous and didactically it's a disaster. I now feel sorry for given Andrew Ng's course on Machine Learning only 3 stars, compared to this it should have received 7 stars.

PS I have 30 years of experience in teaching data analysis and statistics, both in industry and at university and I have 35 years of programming experience.

It was the worst experience I had in coursera. The video content is too general and has little connection with assignment. The assignments are challenging but poorly organized. Students are encouraged to search and self-learning, which is irresponsible and inefficient. I register the course to learn things systematically and save time! I spent about 8 hours on searching desperately for each assignment. My passion for Python and Data Science almost disappear.

The version of pandas used in this course is 0.19.2, which is not convenient for using some functions.

The assignment is not explained after grading. I had no idea about how well my solution is and whether there is a better way to do it. (The forum may be useful, but it took too much time for searching suitable answers)

I have learnt Java before and had little experience in Python. Maybe it's the reason for the longer commitment on those assignments. However, as a former teacher, I do suggest some improvements on the pedagogy in this course and take more responsibility for the students.

Extremely dissatisfied. Lectures are useless and the instructor didn't put any effort into designing a curriculum. After struggling for days on the second assignment I purchased the recommended text for the class. Week 2 effectively starts at chapter 7 in the book. Instructor needs to replace his absurd amount of face time some slides showing the application and logic behind the methods he is trying to use. Jupiter notebook is a tool designed to make the instructors job easier, doesn't help the student at all. I could go on but this course has drained the energy from me. FYI, half the people that complete (or attempt to to complete) this course don't continue on with the specialization. This should speak volumes.

This is not a course, rather just guidance to use StackOverflow. The trainer Prof. Brooks is highly unlively and plainly reads out/speaks some statements. He teaches only 10%, remaining 90% you need to explore on your own. The Assignments have the most difficult questions and for solving them, students are not even given any getting-started questions, to begin with! If you wish to learn Data Science/Data Analysis then I would not recommend this course since it is not worth the time, effort and money. Also, the title of the course is devaluing the efforts we put it. The entire course is focused towards using Pandas to perform Data Analysis / Data Cleaning / Data Wrangling / Data Munging / Data Preprocessing and thus I would recommend that the title of the course should be one of these rather than "Introduction ..." which hardly gives any weight to what hard work this certification demands!

In the brief course videos professor makes some hand gestures and the background shows the people working - both has no relevance and rather prove to be a distraction. Above that auto-grader comes with its own lot of problems which consume hours and days of all the candidates. This course is online for more than 2 years now and I doubt if Coursera really takes such feedback seriously and takes any action for improvement!

Request Coursera / UM to go through all the reviews:

https://www.class-central.com/course/coursera-introduction-to-data-science-in-python-6671

I have decided to drop this course. I have worked very hard on Assignment #3 for the course and can't seem to get back from the grader anything more than "Your solution file either prints output to the console (e.g. through the print() function, or it fails to run (e.g. throws an error). You must make sure that the py or ipynb file which you submit does not have errors. The output we received from the ipthon interpretor was:" That's all I seem to get back.

I'm not sure you should be paying for a course where you have to search the Internet to learn how to do the assignmens. I could be doing that for free! Is the certificate worth it?

This is a terrible course. The "instructors" give quick little lectures, then we're told, for the somewhat complicated (for a beginner) assignments, to look things up on StackOverflow to figure out how to do the assignments. Um, no. I want a class so I don't have to tear my hair out dealing with the internet.

The Johns Hopkins Data Science Specialization was fantastic. They actually taught how to use R to do data science. I'd hoped for something similar for Python with this class, but I guess The University of Michigan isn't Johns Hopkins.

I found this course terribly bad and was the worst experience I had in Coursera. The instructor, rushes through few minutes of video explanations and then it is basically a learn everything by yourself by reading the panda documentation. If I wanted to learn things by myself reading the panda documentation, I wouldn't have taken the course! Like someone else mentioned, I registered the course to learn things systematically and save time!

The assignments require us to do things in panda which where never explained in the lectures, nor it was explained how to do them efficiently. I managed to do everything through hours of research, but I'm pretty sure there are better, more efficient ways to code the assignments but these are not taught, the assignments provide no feedback and the forum too hard and time consuming to search for those answers!

So what I learned was that I could either spend hours and hours cleaning the data store in Excel files in panda or few couple of minutes directly within Excel. Because the course is mainly a do everything yourself, it totally failed to show me the power of panda as a powerful tool for data science.

I have to say, the quality of this course is significantly inferior than the previous "Programming to everybody" . Firstly, it has something to do with the language that the lecturer uses to explain the concepts in Python. The language he uses is unnecessarily complicated. The sentences are very long and the words he uses are vague . If you want to explain a relatively complex concept, you need to use simpler and more comprehensible words. You cannot use a complex concept to explain another complex concept. The second thing is about the structure of the course, the insufficient engagement. The lecture is full of contents but only with very less interaction exercises. This will make students lost in the half way. I have to say this is not a pleasant experience even if I'm already very familiar to some other programming languages such as R and C++.

Very rare I write review. And this is rare course that teaches nothing. The lector explains nothing. "I encourage you self-learning. Read the documentation. Search google. Ask stackoverflow.com. If you don't know python, take another python course. If you don't know statistics, take another statistics course. I encourage you self-learning even more." - that is what lector sais to you.

I would not recommend this course at all. I cancelled after week 2. It is framed as an intro to data science but the teacher often packs too much information into each video without taking time to properly explain underlying concepts sufficiently. The exercises often have issues in datasets that are not linked to current lesson which makes it confusing to follow along and you can spend hours searching forums to find solutions. In addition, the course does not give sufficient insight in videos to help you with questions. You have to search stack overflow and other sources to work out answers. Whilst this is reflective of real life, for an intro course which is how this was framed, it is difficult for someone new to data science to master a concept whilst trying to solve for other non related issues searching the web.

Huge disparity between the course videos and the assignments provided. A background in basic programming is highly recommended. The assignments provided are of good quality however, and provide a great learning experience given one can get through them.

overall the good introductory course of python for data science but i feel it should have covered the basics in more details .specially for the ones who do not have any prior programming background .

The course covers good grounds if you are new to pandas. However the content is pretty much a read-through of documentation and assignments are designed poorly:

1) There are loads of knowledge required to finish the assignment that are not addressed in the course. You will need to find those out yourself online, which in itself isn't a problem but you are also paying coursera for the time when you pretty much do self-learning

2) Due to how assignments are graded, you have to provide EXACT answers as the instructors wanted to see, meanwhile you don't get all information needed on what should be the right answer as certain parameters of the assignment are not clearly defined. The forum is disorganised and you end up spending a lot of time browsing through historical posts, some of which remain unanswered after months

3) Too much time spent on data cleaning. Granted that in real world projects data cleaning is at least 80% of the time but this is a learning course. You don't get to spend a lot of time to practice contents covered in video but struggling with the random complexities introduced by the raw data + specific results being expected. This leads to a frustrating experience.

Suggestion to the course instructor: re-design the assignments to make sure students spend more time on key points covered by video. Have one assignment focused on data cleaning, but not all of them. Also review forum questions and understand where people get frustrated, and improve your instructions in the assignments.

Suggestion to fellow students: it is probably best to enroll the course, note the content/ resources, quit (so you don't need to pay) then learn yourself on google/ stackoverflow. This will be pretty much the same experience.

Excellent material. Admittedly I can see why there are so many negative reviews about the ambiguity of the assessed tasks. It won't be an easy course for anyone who is unfamiliar with programming. However, if you do have programming experience under your belt, you'll likely find this course strikes an excellent balance in terms of conciseness, practice, and theory. Each lecture is crafted carefully to teach you about some nuance of pandas or numpy, and the programming assignments are packed with coding questions that will help you revise what you have learned, in a very efficient way. There is very little "fluff" in this course, which is a major weakness I've seen in similar courses of its kind. Too much spoon feeding often does not challenge or engage the learner. The course is very direct about what it expects of its students. Every week there is a comment "This week's assignment requires more self-learning than the last". And true to its word, there is less and less hand-holding as you go further into the course. I thoroughly enjoyed the material and probably learned the most out of this course than any other course I've taken on Coursera, taking in to account its length.

I thought this was course was good, and was fairly challenging for an online-only course. I thought the lectures could have been a little longer to ensure proper coverage of materials and functions.

So, I have some strong opinions about this course. It should really be called Mastering Pandas for Data Science because despite the word introduction in it's name this course basically starts at an intermediate level and skips straight over teaching the basics of anything, including Pandas. You will have to already know quite well: Python, Pandas, Database style thinking and working, Statistics and Autograder intricacies. It basically took me well over 100 hours to get it done. If I had already learned Pandas or R elsewhere before it would have taken less than half that but basically everything I did had to be looked up on StackOverflow or in the Pandas documentation. However, at the end of it I feel like I actually know basic level Pandas and I am now intermediate level. In the end, a great course but very, very challenging.

There are several inherent problems with this course and how it is structured. I would like to preface this review by stating that I am likely below the level of python mastery required to complete the course; I came in with a simple background of some time spent learning from introductory texts, but I've never worked with the language professionally and I have been out of practice for a couple years. The course is labeled 'intermediate', though how 'intermediate' is unclear since there is only one elementary course recommended to get you "up to speed" enough to take the class. I used my free week as a test to see if my background was enough to be able to hang with this; it was not.

However, even for someone with more depth of understanding than me, this course is poorly designed to teach. Each lecture is between 2 and 10 minutes long, and they do a very poor job of explaining concepts. The notes provided in the supplied notebooks lack any sort of meaningful documentation, and so it becomes trivial to get lost as the lectures go on at lightning speed and you have no comments or notations at all in-text to explain how one thing they did is different from the next.

Furthermore, the difficulty and complexity of concepts spikes up rapidly. Using the in-lecture questions as a barometer, the complexity jump is like going from addition to calculus in the blink of an eye. The first in-lecture question during one of the early videos requests you make a very simple change to a supplied piece of code to add three numbers instead of two. Two in-lecture questions later you're asked to create a list of all possible combinations of a number and letter system for a hypothetical internet company based loosely on concepts you learned 15 minutes ago and never before had a chance to break down or practice in any real detail.

After quiz 1 is where I threw in the towel. Looking at the future assignments, as well as the discussion board, things seemed like they would only get worse, not better. Assignments are highly complex and rely on an understanding of python well beyond what this course, and the one before it, will give you. Users with years of programming experience were having to rely on google, text books, and stack exchange to solve the assignments.

This course is too light on details and too focused on speed. You'll essentially be teaching yourself what you're supposed to do. As my first attempt going through a MOOC, this was a letdown. Part of this is on me, as I do not have the requisite background to fully engage and succeed here. Part of it is on the structure of the course, which truthfully seems to focus on speed and ease for the creators rather than in depth of understanding for the students. If you're considering this track, I recommend instead picking up some textbooks and going through this yourself. You'll lose the certifiable, but you'll have a less frustrating time.

Wow, this was amazing. Learned a lot (mostly thanks to stack overflow) but the course also opened my eyes to all the possibilities available out there and I feel like i'm only scratching the surface!

This course was fast paced but the material was interesting and not to complex. I can only recommend this course to anyone interested in Data Science and who already has a basic knowledge of Python.

My take -- assignments for week 3 and 4 leave a bad impression. I am quitting halfway through assignment 4. I won't take another class in this sequence. I have taken ~10 online classes this year and only gave up on 2, including this one.

I wasted too much time in weeks 3 and 4 cleaning data. When result dataframes have hundreds of rows it isn't trivial to scroll through them looking for problems. Time wasted on this would have been better spent on pandas. I get that data cleaning is a big part of data science, but this is a class not a job.

A few questions tended to be vague on requirements but then very particular on what data is acceptable. The worst is getting a question wrong because the answer has type int64 and the grader wanted int. It doesn't help when the grader just says 'wrong answer' and there are no visible tests. The first question on assignment 4 is a good example of that. I did finish convert_housing_data_to_quarters, but from reading discussion forums that too has a few unstated requirements.

There were also a few typos in assignments 3 and 4. Again, if you want my answers to be really accurate then invest some time to make sure your questions aren't ambiguous.

This could be a valuable class, but you need to invest more in developing the assignments.

I have been doing some coding in different languages and this is my first time in Python. I would say I am on intermediate level (7 years practice), but this course made me spend a lot of time on learning only a tidy bit. The course fails in the basic educational purpose: Provide theoretical lectures and apply it with practical experience. There was little to none connection between these two aspects. The exercises were also way too difficult vs the lectures. I would never recommend this course.

My recommendation: Don't make the student spend too much time googling, and solving problem outside of the curriculum, but provide many easy questions as it is a introductory course. Also, Provide a better overlap of exercises vs lectures.

Some 2 years after finishing this course, I cannot stress enough how much I have gained from this course (or the full specialization for that matter). Having started this specialization as a social science researcher with a solid background in traditional statistical research (and total beginner with Python), I have actually managed to find a job as a Data Scientist halfway through this specialization.

This course and specialization will teach you how to use all the commonly used Python libraries for Data Science applications (Pandas, Matplotlib, Numpy, Scikit-Learn, NLTK etc.). And comparing to some of the other specializations I've taken in the field of Machine Learning, Deep Learning and Maths after completing this one, I can say that the programming assignments are by far the best I've encountered on Coursera so far. The learning curve is pretty steep at first (it was for a total beginner like me) but you'll learn a lot quickly. And by the end you'll be able to do most Data Science tasks independently. Highly recommended!

The difference between exercises and lectures is too big. You end up researching elsewhere more than just follow the course.

Exercises being graded is a challenge > installed grader version is different (older) than the note book. This led to the point that my exercises weren't graded due to exception because I used a more recent feature of Python Libraries

Week 4 programming assignment was a challenge: To pass the final question you need to figure out the previous questions as well, other wise you do not get the right answer. In case you got it wrong it is near impossible to figure out the root cause (because the grader doesn't give you elaborate clues where you might correct things). You do not get enough side information on your own so as to solve any issues on your own.

The content covered in the specialization would exactly be the things I'd want to learn, but the learning experience was full of bumps ...probably I won't follow up on the next courses of this specialization.

I heartily recommend an overhaul of this course, giving learners some more background by more explanation & providing more consistent information to master programming exercises.

On the positive side: Staff was helpful for a couple of open questions,. Even as the learning experience wasn't as good as anticipated I learned quite some Python stuff due to the contents of the course.

The class is mostly problem set focused with no feedback on what makes for good pandas code beyond it getting the correct result. Not a good resource for learning the pandas library. Very little teaching.