AS

:)

-

The information was somewhat confusing at times and it was kinda hard to follow the lectures even though the information provided was quite basic nad not too complex. I guess the problem with this course is the way the information presented and the overall flow of the presentation.

Also the labs, they confused me even more because we get presented with some amount of code which was not covered before. You are supposed to be able to complete this course without any coding, but you get all this unnecessary code, which doesn't even matter in the end but adds to the confusion and makes the lab harder to follow. I think it would be better to get rid of the code, or to include these labs after the python course, so the students can easily follow what's actually going on in the labs.

As i figured from the discussion section there is a number of students that were a bit confused about what actually should be in the final assignment (myself included). I had to rewatch all of the videos and revisit all of the labs just to get vague understanding of what needs to be done.

I am still unsure if what i wrote in the final assignment was even 100% correct (even though i got the top score), simply because these assignments are being judged by peers, not mentors.

The course "data science methodology" provides a reasonable good overview of the main stages of a data science project based on a methodology similar to CRISP-DM methodology. Explanations are supported by two main examples: one related to "reducing risk readmission of patients in a hospital" and the other related to a study of "food recipes". In addition, some Jupyter Notebooks have been developed to make the course more practical. In general, I believe three weeks is not sufficient to cover all the phases of a data science project with enough detail. I think the video recordings do not provide clear and sufficient explanations of the different phases of a data science project. I also believe the course should provide further details, examples and Jupyter Notebooks to better address relevant issues in each of the phases of a project. I would add more "optional content" for students interested of additional details and examples. The final assignment seems not to be sufficient to prove that a student has understood the material. I think this course might benefit if students are already familiar with programming languages such as Python and query languages such as SQL (in other words, please consider adding these prerequisites for the course).

Found that I learned best by reading the video scripts rather than watching the videos. Perhaps due to the fact that the material pertained to a methodology, which tends to be a abstract. The examples were a good start to applying the methodology, but there were a few gaps for me. Just as one example, the Biz Understanding portion was unclear. I intuitively applied my 6 sigma background and BU seemed to be like a definition of the problem statement in terms of measurable metrics, but the script and case study did not adequately walk through. When I read some of the peer assignments, i felt that some may not have applied the concept in the way I understood it. Another example, in Data Prep an illustration of the takeaways / deliverables from doing descriptive stats, pairwise correlations & histograms might have helped me visualize how I would apply the insights gained from those 3 steps to the data.

A more positive example is the illustration of data manipulation for missing data, bad data, etc. Relatively clearer.

This class was OK. Solid introduction to the start to finish process of describing and solving a data science problem. Not super engaging, but that's OK.

My biggest problem is that the step where you turn a business problem into a data science problem is glossed over. I think they called it "analytic approach". It's easily the most important part of the course, and it is given very little attention.

This comes into sharp focus when you try to do the final project, and realize that unless you've done this professionally before, you really don't understand how to ask a question in a way that sets up the data science methods.

As an overview of a method, it's not bad. It really does highlight the iterative nature. But the final project is maddeningly vague and nearly impossible to do due to the "skipped" step inbetween.

I know that they didn't want to teach statistics, or assume people already knew statistics. But then the finally project should have held our hands a little bit better for this one step.

I wanted more out of this class and I think this entire certificate should use this methodology as the manner in which all the classes and projects are done. It was still good to take to get a good foothold of the methodology, but by structuring that same methodology towards this certificate would go a long way in enhancing the overall experience. The first class could go over the analytic approach. The next three or four over data requirements and gathering data. Another three over the exploration and then use the final two classes or so to go over modeling and tweaking. There's potential for such a concept. Make it so!

I felt like there was too much emphasis on a top-down approach. Many a time one doesn't have the good fortune of going through the entire data science methodology as mentioned here. The client has already collected the data and then comes and gives you a problem. In this case, you need to have a bottom-up approach - play with the data already collected and see which analytic approach is feasible. In addition, not enough was done to say that this 'story' is the ideal scenario! Rarely do you get the chance to do a data science project so neatly. But it is always useful to know how things would work in a perfect world.

The actual methodology and the questions to keep in mind for each step are very good, and it's good to have this foundation for understanding data science. But the course was poorly designed and not engaging. Too much jargon was used for a beginner course without explaining what terms mean. There was a glossary in the intro but it was just a list of words with no definitions (were we supposed to look them up??). I'm a native English speaker who works in healthcare and even I felt that the medical case study was too dense to really understand as a case study. The recipe analysis in the labs was much better.

A good course to get insights about methodology used within Data Science to analyze and visualize data to make effective decisions. I thank the professors to make this course interesting.

A couple things which I think can improve the quality of this course. Videos can be made in a better way so as to facilitate people with non programming background. Also the case study used to explain the concepts in the videos isn't the same as the one used in the notebooks. If the case study used is same in both videos and notebooks, It would enhance clarity of the taught topics.

The case-study of CHF and patient admission is an interesting one, however it is quite hard to follow, as much "domain knowledge" terms are used, which are not really familiar to everyone. A glossary of terms or some explanation could be useful. Also, many times the slides are separate from the talk i.e. it is hard to read and listen at the same time. For example, many times a "cohort study" image is shown, which is not explained entirely.

The cuisine lab training with python code is overwhelming. I don't see a reason to view a code, without knowing to code.

I appreciate the objective of this course - to provide a high level overview of a very complex process and to emphasize the role of the business requirements, feedback, and iteration in the data science methodology. As a relatively concrete thinker, I would have benefitted from a more practical application to a very small problem (perhaps one that could be performed with just excel). This could emphasize the context and feedback portion without requiring the programming background.

A very good overview of the problem solving methodology for data science projects. The capstone exercise was practical and helpful to put all of the pieces together in a logical order. Perhaps analytic approach and model development and deployment could have used additional modules or case studies. The single module for each is a good start, but a second case study could better illustrate the difference among predictive, descriptive, or prescriptive approaches and outcomes.

A subject that can take a whole year was packed into a few short videos, if the goal was for me to have a vague concept about data science methodology then maybe the course is ok. If they wanted me do really undertand the methdology then they did a poor job |I might as well study this by myself searching google, they just gave headlines of the subject , this is confusing I hope the next course is better because this and the previous one are under par for me at least.

This was the first course in this series that seemed to provide some knowledge. That said, the cognitive class external tool is painfully slow to use. I'd recommend skipping the ungraded assignments, as the payoff isn't worth the time you'll waste waiting for the notebook to open. This must be an obsolete tool, which IBM stopped supporting at some point.

I'm hoping the next course will allow me to run python on my cpu, rather than using a broken cloud tool.

The course is good and interesting, but I feel that it lacks the hands-on part, and that it could be more engaging. I feel that this course should be after the students have a tool that they can manage the data with, and that they can start dirty their hands with data.

The course as a stand alone course doesn't contribute a lot - it's interesting only as part of the whole certification, and should be linked to other tools in order to bring more value.

Lab exercise/further reading doesn't make sense to me since I'm new to data science. Got a headache following what happening with the codes. The methodology introduced here is an IBM methodology and its pretty easy follow. Some of the terminologies are not enunciated clearly and it's pretty hard to track and understand. Overall, this course is a basic understanding of Data Science approaches and the use of important use methodology.

Overall l it is a very good course. but on the lab section, the instructor's english is not very good. He can not deliver his thinking very well. You have to translate it to your self everything you read on the lab. In the business understanding section, he can not deliver the problem. Readers can not understand what he wants and what the goal is. IBM should rewrite this section so make it easy for readers to understand it better.

A little wordy with the labs focused on shift-entering prewritten code as opposed to giving significant input. Also felt that one peer-grade being factored into final deliverable is a little sketchy. I had one peer completely fail my deliverable selecting lowest marks on each section of the schema yet when submitted a second time with no change (i was honestly happy that my deliverable met the requirements) i was awarded 100%.

I think the content presented was okay, and was generally presented quite clearly. The labs were well structured and easy to follow, but I didn't feel that I was learning skills to understand when to use different methodologies, or what kinds of challenges I might face along the way. The example given was clear and easy to follow, but I don't feel that I learned a lot that prepared me to analyze other data science questions.

Very general and abstract presentation of what the Data Science recipe is. Still nothing practical three courses into the data science specialization... Had I followed the schedule, I would be 9 weeks in with nothing to show off. At least, this course gives a nice overview of what a data scientist will be doing, but I think this should have been presented in the first week of the first course, without necessarily testing it.

The exam for week 2 was terrible. The questions were way too tricky it was not necessary. Also I only was reviewed by one peer for my final assessment. This was bad because I deserved 100% and they gave me a only "Good" mark on one section probably because they figured out I gave them a "Good" mark on a section which they only did good on. More peer reviews should have been done than just one. I deserved a higher grade.

An informative introduction to data science methodology, but the presentation of the material could use more work. The videos could use better production values, with perhaps a bit of music and more visual aides. There is also an annoying six seconds of silence at the beginning of each video which made me think there was something wrong with my audio. It would also be nice if some of the labs were a bit more interactive.

On overall IBM data science professional certificate track: Pros: Content is just good enough, instructors are good. Cons: IBM watson and the platform given to practise on is awful and has terrible performance and reliability issues, most often doesnt work and had an impact on my test deliverables. I personally overcame those issues to some extent with kaggle's and google colab jupyter notebook environments.