I found this course useful since I had no previous experience of Pandas or the statistical features of Python. I have programming experience and feel I would have found it quite a challenge otherwise, but not prohibitively so. Chris is great at explaining things in an accessible manner, and I'm very much looking forward to going into more detail in the rest of the specialisation. I would have enjoyed a little more focus on statistics etc and a little less on the mechanics of the library, but understand why it had to be approached this way.

For future tech or CS courses from U Mich, please follow Dr Charles Severance's approach in teaching. Make learning and wanting to learn truly enjoyable. Teach material that will be tested in assignments and not throw curve balls which lead students to google answers and 'stack overflow' it unnecessarily.

This course is not for beginners or for those that have just picked up programming.

The autograder system is a real pain in the butt. A lot of time is wasted just to conform and present the answer that the autograder will accept.

Great course material. Grading issues due to Python versions was a pain. The discussion forum may have been more helpful with smaller groups (18k posts in a given week is overwhelming!) Nice to have access to the larger group brain, but I bet having chats with cohorts of 10-100 people would make the course feel more intimate and engaging. Also, the pace of lectures felt like the prof was reading text off a screen rather than a "real" lecture with natural pauses, which made it difficult to follow along with code as the prof was speaking.

This is course is classified as intermediate and a 10h commitment per week. For me this was an under estimation. It took me much longer. The problem with the course for me was, that the skills required for the assignments were not taught during the lectures and the assignments are poorly described. I ended up googling for hours. The links in the forums are very helpful, but I usually prefer to solve assignments on my own, which was not not possible for me in this course. I did learn a lot, but this was very unefficient.

Honestly, my opinion is that the material is "a little too difficult a little too early" .. not because an experienced programmers can't handle it, but because the urgency of getting the final assignment done forces a reversion back to more traditional programming techniques rather than instilling "Pandas like" programming techniques ... if anything .. "instructor solutions" should be given at the end of the course so we can go back and see "how we could have done it more elegantly" in the areas were are interested in.

Not a bad course but would like to see more teaching of best practice solutions to some of the test and assignment questions. As most of the assignments require a lot of self-learning it would be nice to discover if our solutions are optimal or not. As it stands you can get a perfect score by writing for loops or other inefficient solution when there are quite possibly built-in pandas functions which could achieve the same thing more efficiently. Would like to learn more about pandas and best-practice techniques.

The course materials is very practical. The lectures are very clear and self-contained. The only reason I gave 3 stars is that the homework takes too much time. I spent a lot of time digging into online forums to find out the nuts and bolts to finish the projects. Fortunately the teaching staffs are very helpful. The time spent for homework is too much for my preference. Maybe this is the way to learn this type of information. I don't know. Having said that, the materials and lecture qualities are great.

I liked that the course was very assignment-oriented. It had a good structure and interesting additional readings. However, for people who are not familiar with pandas library it may be very challenging to pass assignments. This minimal guidance, I believe, results not in a better understanding, but in confused students writing bad code and spending hours not understanding online documentation. Overall, I really liked the idea and content of the course, but not how instructors approached the self-learning part.

I have two general comments:

The first comment is . . . there was too wide a gap between the lecture content and the assignments. The second comment I have is . . . I spent too much time trying to figure out what the autograder wanted and not enough time learning Data Science with Python. I can only imagine the work that it takes to develop and launch a course like this. In all, I'm very excited to be part of this program. My comments are critical but hopefully helpful and all your work is appreciated.

Course content was generally good although sometimes the lecturer brushes over topics that could do with more explanation. He may explain 10 things you can do in quick succession making it hard to remember all of the points. Exercises were good but there is a big jump on week 3 and 4 and relies on students to spend time themselves searching the web for solutions to the problems. Whilst it is good to be independent, asking for things that haven't been taught in the class can be hard for some.

This course teaches some useful techniques, but suffers greatly in it's ability to teach you those things. The hw's do not contain enough information, and submission errors are very vague in terms of explaining what is actually wrong. During lectures, too much time is spent on the presenters face. They have presentation slides and code examples that do not get enough video time. It is not at all useful to talk about techniques without giving any textual info about it. It's not easy to follow

It was a good course where i learned about new and great tools and techniques. I learned how to approach the data science problems using Pandas and Numpy. This would not serve as a great course for into to data science. Background with Database Management and Python really helps. Overall i learn about new and great tools and would definitely require Documentation while using the skills i learnt in this course. Overall Great Job by Professor Brooks. Would love to take more courses by him.

I have reasonable amount of experience in data analysis using R. I found this course to be fun and challenging. But I heavily relied on my previous experience. I imagine it would be very difficult for someone with just basic Python skills. Also, it feels like a large chunk of assignments was about data cleaning (yes, it’s important) while important pandas concepts weren’t covered in as much detail as I would’ve liked. The recommended book is really good, I’m using it as reference a lot.

I enjoyed and found all of the lectures helpful, but lack of feedback after submitting assignments was a real problem, especially for the last assignment. A simple response of "this was answered incorrectly, points not awarded" isn't very constructive and was often frustrating. Bugs in being able to submit the assignment were also frustrating. I spent a good amount of time trying to fix my code, thinking it was incorrect, when in actuality the online submission was just not working.

Course was good and i have learned and starting to refining my skills because of this course.

We have to do lot of practice and net browser to work on the assignment and only think that lacks in this course is the study material in the form of video i think,if the content of the video increased then this will be the perfect course plus there should be two assignment one related to what you have studied during video content and the programming assignment which test your skill test.

I have mixed feelings.

1) the course is very interesting. It is an applied course, there is a lot to learn.

2) But very hard too. There are so many things, that you may have difficulties to memorize quickly through practices. Then, you end up roaming on stack overflow and pandas documentation because you just can memorize enough to process and have to rely heavily on the documentation.

If you are interested to dig into the python, this is very interesting (but quite hard)

This class definitely makes you learn, but not as much from the lectures and course materials themselves, as from the discussion forums (shout out to teaching staff and mentors for their great help) and online tools like Stack Overflow.

I understand that this is also a technique to make us more independent, but it seems like the professor just wanted to skim over this part in order to concentrate on some future / more advanced class that is more interesting to him.

Although I learned a lot in this course, I found the lectures and assignments to be much too different from each other. I would like to see assignments where you must practice what is learned in the lecture. For myself, I feel I learned 1 set of concepts from the lectures and another set from the assignments by spending time on stack overflow and the pandas documentation. Both are good but the lectures and assignments did not "flow" together as I would have liked.

The course is very interesting. The Jupiter notebook is very useful.

I don't like that many examples are very US-specific. Some important terms may not be clear for people who live outside USA.

Update: I'm lowering my grade from 4 stars to 3 stars because of very confusing assignments. Often it's not clear the requirement of the task. It takes very long to understand. Also some assignments require methods and functions that were not covered in video lectures.

course content is good,but the instructor tries to explain everything just by saying it. does not demonstrates it every time and does not dive deep into the concept,so that the learner may get more interested. and if you are stuck at any assignment question then it might happen that you won't get the answer even after seeing the videos. assignment contains some questions based on concept which are not discussed,so you have to figure them out yourselves.

Excellent lectures and explanatiom about Pandas features. But the Assignments could have more conection with the lectures of the correpondent week. Also the instructions to code the assignments are often bad written or lacking information, causing erroneus comprehension about what are being asked. These kind of problem cause a lot of misconceptions abou the task and cost a lot of time from the student just to really understand the assignment objective.

The assignment of week 4 is wrongly explained in the jupyter notebook. It says that the price_ration is equal to quarter before recession divided by quarter bottom recession. When you do so you will never get a validated result. The only result validated is recession_bottom minus recession_start!

Giving assigments should include expected solution. Now you upload your file several times in order to figure out you are using the wrong formula.

In my humble opinion, this course does not have a correct balance of difficulty of new concepts/tools with the exercises given by the course to practice. It is a "hey, look, there is something called pandas out there that may be very useful for you" but it is too introductory. And I know the course has the word "Introduction" in the title but that does not grant, in my opinion, to run a 4-week course such as this one with so little content

The whole curse relies on the automated grading system, which is still a little sloppy. I think it would be useful to have one notebook or cheatsheet with all the important functions. And also, I personally like to have a reference assignment that was completed by the instructors. This could be provided before or after the course, but the way I really learn is by reading other peoples code and seeing ways to code problems. Thanks

A good introduction to python and data science. The questions were just about the right level of difficulty. My main criticism is that the online videos were pretty short and not going into a lot of detail, whereas with the questions you had to do a lot of extra research to figure out how to solve them. More interaction with the enrolled students during the course and having more in-depth videos would make the course a lot better.