Very basic

Too basic.

it is good

Too basic

too easy

Too easy

OKKJUGH

It's Ok

Shallow

hosome

爬虫部分很烂

ok ok

great

Good

good

good

good

cool

good

ok

I was very excited for this course, but felt underwhelmed overall. I have some programming background, so expected the course to seem a bit slow having been introduced to topics such as basic data types in Python before. Week 1 material, for example, was simple enough to go through in less than a day. As the course progressed to more complex and interesting topics (e.g. APIs, Webscraping, etc.), however, the course seemed incredibly rushed. Instead of understanding the set-up, it felt as though the lectures simple flashed multi-lines of code. The labs similarly seemed to grow in complexity, but shrink in explanations. The labs often incorporated libraries or methods/functions that had not been referenced in the lectures, often with little added explanation/introduction. Similarly, the incorporation of the Watson engine seemed like unnecessary advertising. It was not necessary to understand Python for data science and AI as part of the exercises. So much time was spent explaining how to add arrays, while it took 4 minutes to go through summarizing dealing with csv, JSON, and XML files. It also seems that later lectures had more typos in slides, again giving the impression that the later sections were rushed. Even the final lab randomly referenced a library for processing image files, which seemed to come out of left field. The course could've stopped at Week 4 or maybe split the focus between pandas and numpy between Weeks 4 and 5. Or more focus on exception handling or classes and objects might've been useful. Use of Jupyter notebook is helpful (it really adds to nice-looking labs), but I think it would've been better to have files/resources that could be downloaded and saved locally. That way, any files written/read could also be inspected more closely. Given the title of the course (and IBM as the provider), I would've expected more. With the exception of the weak Week 5, the rest of the topics seemed more like an intro to programming (using Python as the language of choice), rather than a focus of Python for Data Science, AI & Development.

This is a course useful in getting you onto the ramp to Python and the Python libraries "Pandas" (provides tools for data manipulation and analysis, in particular the "data frame" matrix as known in R) and "NumPy" (provides multi-dimensional arrays and numerical computing). It also explains basic skills in loading and writing files in various formats from and to disk, getting pages from the Internet, how to perform client-server communication in the REST approach and how to scrape websites. Sadly, it feels like a "0.8 version" of a proper course, or a course which has been abandoned. Content is affected by confusing text, typos, omissions of important points, code in "non-pythonic" style, in some cases code that doesn't work, or looks like it deserved a refresh to latest notation (the matrix multiplication with the '@' operator isn't even used). Sorely missing are chapters on "gradual typing" in Python, assertions and unit tests, all of which I consider a minimal must-know. On the other hand, the "coding notebook" practical exercises (based on Pyodide running in your browser) are the right way to go. Drilling into details and looking up documentation on the Internet is unavoidable, indeed it will give you the understanding that you actually need. The course seems also unsure about whom it is addressing. I doubt that aspiring data scientist need extensive explainers about how vector and matrix operations work, those should be assumed known. Still, all of this should be easy to fix by IBM in a fortnight (I left them extensive notes, too). On the other hand, the questions that Coursera throws at you in the "graded exercises" are rather embarrassing, and need serious rework, I want more interesting question than multiple choice questions about syntactic details, word definitions and Python one-liners (which cannot be written in Python btw.). Make it better, guys!

Generally, the course is very informative and useful.

However, the Python for Data Science and AI course is anything but helpful.

The course requires the students to open a free IBM Cloud account, which is practically impossible in my case for unknown reasons.

The IBM Cloud service is essential part of the certificate program, as it is used in most of the courses, but, however, impossible to use.

IBM Watson Studio is a tool developed for Machine Learning and a part of the IBM Cloud service and an essential part of the course, but can not be used without a cloud account...

I think the above sentence says it all.

Through threads in the forum I requested a solution for my issue, which many other seemed to have. I received a few answers to my inquiry, suggesting a few workarounds like using a non-generic e-mail or at least a GMail, changing my network settings and so on.

None of them seemed to work.

The solution that was suggested at the end was to download the file from a suggested link, run the code locally and upload the result to GitHub...

Seriously?!... o.0

A few e-mails sent to IBM Cloud support remained unanswered.

The saga above, as well as the lack of exercises, non-working code in the lectures and LAB really made me question my choice of certificate...

The course does not contain much more information than the one you will find in the book by Murtaza Haider, parts of which are included in the reading sections of the course. If you are wondering if it is better to spend $35 each month on Coursera, or to buy the book for $20 and learn the same stuff... Well, I think you have your answer.

I would not be recommending the course to anyone, as I am not sure If they will cope with the frustration around the process of finishing it.

Sometimes I got freaked out when I completed a course with flying colors and still felt like I didn't understand anything. Unfortunately, that's exactly my experience with this course, Python for Data Science and AI offered by IBM.

If you are attracted by the title, or by the name IBM, then I advise you to stay away from it. There is nothing remotely data science or AI in this course. It should have been called Introduction to Python ,or Python for Beginners, but there are much better courses on this topic, such as those offered by the University of Michigan.

Here is why it is bad: All the videos feature the same robotic voice rushing through basic concepts as if someone is just reading from a textbook. I wonder if the narrator is an AI, not a real human. I could have overlooked that if the content is actually good. In contrast, its scope is very narrow, even for a beginner course. There is no walk through of common Python challenges and mistakes to solidify the concepts. and again, you are going to pick up more Python knowledge from other courses.

Worse, the exercises just promote rote learning and the ability to use IBM's products. There is just not enough practice exercises to help learners understanding. Most of my time spent on this course is actually dedicated to figuring out how to set up IBM Studio, or whatever it is called. I doubt anyone is going to benefit anything from this course.

I've always admired IBM's achievement, but it takes another skill set to deliver a good online course.

What the course covered isn't bad, but the presentation is far from polished. There are many errors in the text and in the video narration, and the videos aren't well edited. The errors are mostly small, but they disrupt learning and give the impression that the course was written in a rush, without sufficient proof-reading or testing. In addition to typos (including in the video scripts), the logic of the course also isn't great: methods and JSON, for example, are both mentioned many times before they are explained. Methods relating to specific data types are discussed in week 1, but methods as a concept isn't introduced until later. JSON data are referred to a lot, but it is only in the final week's lab that JSON is defined and explained. There are also long-reported bugs in the lab interface, which are not adequately investigated or fixed (reports of being unable to share to Github Gists since early last year, but the only responses to these are workarounds or "it works for us"). None of this inspires me to try other IBM offerings.

In short, rather than teaching students to code in Python or analyse data, this course is more of a "taster". It gives you some idea of how Python accesses and handles data, but it lacks the depth and practical components to really give a student the skills needed to tackle data science problems - to even know where to start. There are way better ways to learn Python, and I presume way better ways to learn data science.