Outdated and poorly taught specialisation. My best experience on Coursera has been Andrew Ng's ML course and maybe it raised the bar too high. But that was a course that taught the student (keyword: taught). This does not even come close. Would not recommend.

Started with no programming knowledge, took this while working a full time job, it was not always easy, but i am really glad that i took the decision to go out from my confort zone. Today i can handle topics like machine learning, data analysis and visualization with python. thanks for the IBM team who really has done an amazing job on this course.

Must take to complete this wonderful specialization.

You will have a change to apply everything you learned. And you have the freedom to choose the topic that you are interested in.

After this course, you will have a report, a blogpost and a notebook with complete code. With which you can showcase your achievement along the certificate.

A considerable part of the required work (and grading) is on material not taught in this course or the previous ones in the Certificate (web scraping, turning a Jupyter Notebook into a presentation or a report). Instructions are often unclear ("explore..."). Some technical problems are not mentioned, leading a time-consuming researches, even though it appear in the forums they are known to the teaching staff (using nbviewer to make Folium maps visible in Github).

Overall, it is disappointing to spend so much time relying on Google search when one was expecting content taught and delivered by IBM.

It would be great if resetting deadlines wouldn't erase all work/grades previously done/achieved.

Think this one should have been marked by the course moderators and not fellow students.

To much focus on Foursquare API

This course's content is out of date. Students have to rely on the posts of other students to work around issues with the course. This is a real shame as the other courses required to receive the certification are well maintained.

The course project was changed 4 days before the due date. This is unfair and it is a waste of time. If you don't want to waste 2 weeks of your time I would advise you not to take this course.

This course is full of bugs (outdated) and lack of explanation for certain matter is seen in Discussion Forums.

more exercises, more projects.

If I have to say one thing about Coursera or IBM Data Science Professional Certificate course, I would say it as a Fantastic thing happened in my life, I am so happy with it, and I am not going to leave Coursera for ever.

Very difficult to manage the scope, but it is a self-learning process. Recommend extending the Capstone course another week or two, to encourage the students to go all in on their work.

I finished all he 9 courses in this specialisation. Therefore, this comment basically applies most of the 9 courses.

The video contents and the practice exercises are very good and on point. Instructors are great. However, there are serious problems with the assessment mechanism and this is the reason why I am giving a 4 start.

If you work hard on the assignments, meaning that you study and research well to understand the code, you might end up getting a low score on assignments. This is because of peer-graded assignments. Your work might be graded by someone who doesn't understand the material as much as you do, or even someone who submitted a blank file just to see others' work. You rarely get feedback for the missing marks. As a simple example, I once submitted my work and received 4/11 with no feedback. I instantly re-submitted the same work and received 9/11 from another peer.

Another problem is that you get to see the rubric only after you submit. Some assignments are not clear on what the specific expectations are. The rubric must be clear before you submit the work. Even if you try to be flexible in your solution to address the vagueness, the peers may not show the same flexibility although you do the work properly.

And finally, the biggest issue.. Plagiarism! When I say plagiarism here, I mean copying someone else's work line by line all the way. It is utterly disgusting that it is more widespread than I initially thought. Such cases have been posted multiple times in the forum. I encountered at least 2 cases of plagiarism. The only thing you can do is to flag the submission, but obviously this doesn't stop anyone. What's worse is that those people who plagiarised someone else's work line by line get to peer-grade your own work.

Assessment section of these courses is a mess and has to be seriously re-evaluated. Peer graded assignments can be accepted to a certain extend but not for assignments that require hours and hours of our effort.

To much focus on the use of the Foursquare API, which is outdated is bit. Other techniques learned in the program are not used, only clustering with K-means. On the other hand, you are forced to experience hands-on reality, when things are just working different then expected (which is meant to be possitive feebback). Week 4 and 5 take a lot of time (far over expected 30 hours, more like 60-80).

The content is good but I felt that the final deliverable is too long (47 slides where pictures from 7 notebooks need to be patiently copy paste and query results provided by screenshots..)

I enjoyed working on an open ended project, which was not the case in the remaining 8 courses in this specialisation. I was completely unprepared for some common challenges with a real life dataset and it took me a hard time to address them. On the other hand, it is hideous that only one person reviews your assignment and there are absolutely no requirements regarding the accuracy of the results. I mean who would accept a statistical model with the lowest possible accuracy in your new dream job? Anyhow, the only reason why I kept improving my project is just to show it to my potential future employer and not because of the requirements. The other funny thing is the cavalcade of "Please review my assignment" threads in the discussion forums, which makes it impossible to have a meaningful discussion there. In conclusion, I did other courses on Coursera, but this one had a far lower quality than those.

pushing ibm watson studio too hard, i lost all interested in this course.

This is really challenging course, especially that you get hint on how to use a RESTful API (of Foursquare), how to create heat maps, or create different marking on a map using folium. The Capstone was really challenging, because you can practice what you have learned during the courses of the specialization, like how to start from the scratch a project, how to apply the data science methodology, like business understanding, gathering, analyzing, and cleaning data (most of your time you will spend on this), applying the right machine learning algorithm to solve the problem (modeling), using Jupyter Notebook on IBM cloud and using github. In the end you should also prepare your final report including the business understanding, describing your data, presenting your result, and placing a discussion section in the end. It took me 4 full days to complete the capstone, but I learned a lot.

This is the best part of the specialization and I learned a lot in this Capstone Project. If you've been doing all of the 8 previous courses, believe me those 8 courses are nothing compared to this course when it comes to putting time and hard work. You will learn a lot of things including web scraping, connecting to a url, using geolocation services to get data about a location. You'll also use foursquare API to get popular venues in a particular location. This project is super interesting but at the same you have to put in a lot of work too! It took me more than a month to do this capstone alone but it can be easily done in around 3 weeks if you're dedicated in completing it.

Good class, very useful. Peer grading is a great idea, don't like the practice of posting notes in the forum with subjects like "You grade mine and I'll grade yours." At the least, it gives the appearance of cheating. It is also wasteful, as it leads to some assignments being graded multiple times while others are waiting in the queue. This is a practice that Coursera encourages, which is baffling to me. Even in the last class in a 9 class series, I ran into people submitting blank or nearly blank assignments, with no content or inappropriate content, who were apparently hoping for a pity pass or cheating.

This course is really good and give enough challenge on the final project, especially on how to get data from multiple sources: scraping data from web, call APIs, and visualize it on map after call the clustering algorithm. I like the way we should prepare all material to complete the course like visual presentation with slide/blog post, report, and share the code in GitHub. Really glad I take all these 9 courses, can't wait to take other specialization course.

It was very good.. Overall. few things I like to add.

Sharing my notebook from Cloud was not working a lots time.

GitHub , or Jupiter Notebook with simple 2 lines of coding did not work..

Also, a lot of time, cloud machine just spins.. -- without showing it

My resource got close to limit , - Could not add more code..

Instructions were out of date, could not be applied current version

Working with Cloud machine was challenging !!

Very Good. This is my first contact with data science with python and associated packages.

In the end of the course I'm able to deal with data using python and a lot of tools that helps the job and let this job more fun.

A very well organized and balanced course with videos and very good material for practical labs.

I have a Swisse Knife with me to deal with future researches on data science.

Thanks.

The course was very helpful in presenting me the world of data science, what exactly are the things we need to be proficient in to excel in this field ! Best course of all was Machine learning with Python, you will enjoy doing it ! and we need more questions in quiz to test what student has gained at every step.