The course taught me a lot, but it felt like a lot of self-teaching. The course materials, especially for the final Bayesian unit, were full of errors. It's great for keeping you on a course of study, but it's essentially self-taught. The assignments can be very confusing. For instance, on the final peer-graded assignment, there was a question in the scoring matrix that wasn't asked in the assignment itself, so almost no one did it and couldn't have gotten points for it. The peer-grading in general is a bit of a joke and because of the way the scoring matrix is set up, I was essentially forced to grade excellent projects about the same as very mediocre or bad ones, just because the scoring is mostly a checklist of required components. Lastly, I was a bit disappointed with the final certificate. It is the exact same as the certificates that you receive for each of the sub-courses within the concentration. To an outsider (employer, colleague, etc.), it would be entirely unclear that you finished a whole, months long course as opposed to just one of the courses. I may put it up on my LinkedIn, but I almost feel like it would reflect poorly on me professionally because it looks like a silly online class certificate, so I may just take the knowledge I gained and leave it at that. Glad I slogged through, though!

This section guides you through performing a complete analysis. Some of the grading criteria are inconsistent with the tasks stated in the project. However, in my experience giving the kind of report that you'd generate from R markdown to a non-technical, non-statistically minded stakeholder will result in bewilderment. While Rmd is a good format for documenting your approach and creating a reproducible data analysis, the report presented to stakeholders should take a much different form. In writing an Rmd, you are necessarily constrained by the need to write code in a particular order -- that order may not be the most beneficial one to present. Furthermore, having long summary output of various models and other such code is generally not helpful to non-technical stakeholders. I understand that not using Rmd would make grading more difficult, so I'm ok with Rmd being used here. But I think that it would be good to explain this all to students.

I think this is a very advisable course as a whole, The capstone offers a good occasion to put into practice what has been learned during the four previous courses and also works as a sort of review.

I will recommend to give the information about the projects even if the proect starts in September 10th so we could advance while enrollment opens up.

Thank you very much for teaching me all the statistics courses in the specialization. Although it is an online study, I think I benefit a lot from the course contents, quizzes and problem sets and the training from final projects one by one. Also, I appreciate a lot the feedback from other students in the courses, which gives me more confidence on my study. As a matter of fact, I know I still have a lot of room to improve on my final project, such as consistency of EDA with the later part of my project, assessment on residual plots, real understanding on the trade off between model interpretation and prediction accuracy, etc., which I will improve in my future study and analysis. No matter where and what I will study in the future, I always bring the statistics knowledge and R skills that I learned from my first specialization from professors at Duke University. Thank you.

This was a highly challenging sequence of four courses and capstone, but worth the considerable investment in time and effort. Overall, with just a few glitches here and there, the specialization was superbly taught and well-designed. I thank Mine and the rest of the Stats team at Duke for offering it. I would also like to see Bayesian Stats made perhaps into two courses. Or even a follow-up Bayesian course perhaps not required for the specialization. Professor M. Clyde is one of the world's leading experts on Bayes, and the creator of the BAS package. All respect! I would like to have more of her teaching, but at a slightly less steeper learning curve. To my fellow students and the Duke team, thanks again.

Sometimes capstone courses are focused just on a final project. In this case, the capstone also includes quizzes which force you to review material that will be helpful in doing the final project. Overall, I would say that I felt a little behind compared to my peers in this specialization, since I started from zero; but on the other hand, I was still able to successfully complete it, so with a lot of hard work it is possible.

The capstone project is an excellent way to utilise the skills and knowledge learnt from the previous 4 courses in the specialisation. I learnt a lot about conducting a data science project from start to finish and it allows for ample time to experiment with different approaches to accomplish the project goals. This capstone project, and the 5-course specialisation, is highly recommended.

Very well coordinated Capstone project. A lot of guidance was provided and a progressive review of data handling and Bayesian concepts was made possible through a series of quizzes before culminating into the final project. Really well thought and planned!

This Capstone has giving me a good understanding of statistic's, and the role this discipline plays in the real world. A special thanks to Mine Çetinkaya-Rundel I feel that she does an amazing job communicating the concepts.

Wonderful capstone project, very well guided. The quizes are extremely fun and equally helpful in the completion of the final project. Relays a number of commands and functions with ease and simplicity. Highly recommended!

A well structured course that after having virtually no experience with R or statistics, I now feel very confident at taking on regression projects, and to implement them in R. I found the BAS module to be very useful.

Great course, learned a lot and got me started on another project that I've turned into a really nice portfolio item. I feel much more comfortable with R and statistics principles.

A very useful and practical course. The learner had an opportunity to practice the skills he/she learned in the previous course and use it in a real-life situation.

Amazing capstone. I've learned many things in this specialization. Just try to change the sudden frantic pace for Bayesian Statistics and it's gonna be perfect!

Very interesting project, a great real-world application example. Highly recommended to anyone that finishes the previous classes.

Great lead up to a final capstone project, in which you test a model for overfitting and validate on out-of-sample data.

Challenging and rewarding! The project is very fun and it will prove the knowledge acquired with the previous courses.

A very well organised course providing in depth understanding of linear regression and modelling using R programming.

Great, comprehensive way to utilise all of skills obtained during this specialization!

Very interesting project oriented and practical case! Well done!!

The Capstone project really helped tie the program together

It is great. You really have to apply what you learned.

Great project, real life application.

Strong Capstone Project