Ratings and Reviews for Understanding and Visualizing Data with Python

Good course but definitely wish the practice material was a little stronger or more challenging. I quite like the lectures and the professors and teaching staff definitely know their stuff being UMich's Stats department of course the content itself is great. The lectures are great, the solution sets they give are great but how exactly they did those solutions... well let's just say I personally wouldn't just rely on week 1s coverage of the basics to get to there. I would strongly recommend people have at least a passing understanding of Python like through the Python 3 Specialization from UMich or Py4E from UMich. AND I would say this shouldn't be the first time you use numpy, Pandas or seaborn. I would suggest going through the Numpy Tutorial on the numpy site, the Pandas tutorial on the Pandas site and follow up with Kaggle's micro courses on Pandas, seaborn and data cleaning. This course, true to its name of the stats specialization is really an application of basic descriptive statistics like for Exploratory Data Analysis done with python. Which is what I was looking for so this is exactly what I wanted. Again lectures solid and the solution to the exercise notebooks are GREAT. They don't explain in great detail besides linking documentation how they got there so knowing Pandas indexing, shallow/deep copy, the pandas stats functions, Pandas pivots like melt and stack etc. This really takes someone who knows the basics of Pandas, teaches them the very basics of stats like stuff from high school early college, and applies it to a real dataset as you would in an everyday EDA setting. And it is EXACTLY what I wanted to teach that. Just wish there was more practice on this stuff. Youtube tutorials don't go as indepth imo.

It was irrelevant and contained unnecessary content. Why are we drowning in theoretical statistical topics instead of focusing on Python? Thus far, the course has been more about statistics than actually working with Python! I am here to address my statistical needs using Python, not to become an expert in statistics. Unfortunately, this course seems to be doing just the opposite.

Great course

Good statistics content, but it is not interactive and the testing is weak. Python learning is extremely unforgivable. There are no step-by-step videos, and no theory explanation either, which makes understanding python syntax and functions (particularly in the context of data science) extremely difficult. As someone with an advanced java background, I expected the python learning to be smooth. Unfortunately, I was thrown into the deep end with no life jacket, as the course went from basic variables to creating scatterplots and manipulating datasets in less than a day. This wouldn't be as bad if there were video instructions, but there are none. The "interactive labs" are not interactive, but rather, are just vague notes that don't truly teach or test you on anything. After completing week 2, I left with nothing other than 5 hours of wasted time.

University of Michigan is always the best

It was Perfect.

The courses is supposed to help students learn how to use Python to understand and visualize data. However, the course lacks focus on the subject as well as tasks for practicing Python code. Lack of practice. The peer-reviewed tasks are hilarious - you will be asked to describe how you'd visualize metrics in (Python you would think? No!) words. This is so easy to turn this task into something actually useful: create a notebook with preloaded data and ask students to come up with metrics and visualize them. No-one came here to practice English writing skills, and this shows in the tasks of the students. The quizzes are easy, the final quiz has all answers in hints which are not even hidden. That's actually a pretty good representation of the course creators' confidence in the students' knowledge after the course - we know you didn't learn anything, so we will just give you all the answers. Concentration on the course goal. The course is too short for trying to pack all the information in it. The last week was interesting, but if I wanted to learn about study design, I'd take a course on Study design. A lot of topics can be described as 'Understanding and Visualizing Data', and the difference between a well-designed course with thought-through structure and this course is that the good course is focused around the narrow subject (e.g. using Python for understanding the data) and delves as deep as possible instead of throw in different topics that are related to 'understanding data' in such a short course. And one last thing I would like to bring up is the students teaching in the course. I understand that it was probably the project they got credits for, and the professors thought that it's be a great practice for them. This is a great initiative, but the Coursera students actually pay for this course, and, I am sorry, but the students lectures were bad for the most part - the explanations are not coherent, the repetitions, the 'we are not going discuss that' (then please structure the lection the way the you don't use the function you don't want to explain). While it's understandable that students need more practice in teaching (they are students after all), the question arises as to why one should pay to listen to their 'end-of-the-course project'.

Aweosme Course , Thanks for every instructor , all was great

Generally, it is great! It would be better if examples were provided to illustrate the concepts of confidence intervals and hypothesis testing in week 4.

Great. Lecture-based where concepts are clearly explained and then demonstrated through notebooks. It doesn't get much better than that.