hw06 ready for grading

[hannahdxz]

https://github.com/hannahdxz/STAT545-hw-Dong-Xinzhe/blob/master/hw%2006/hw06.md

f4f8d71

[xinyaofan]

Hello! Hanna, I am really happy to review your homework and I think you did a really great job!

There are some really good things in your homework:

• Your homework is in a logical flow and the format is tidy and clean. I really love the abstract of each part, I can learn what's going on of each part very easily. Also, you give a table of content in the beginning, so I can navigate to each part quickly. Thank you!
• In the first part of your homework, you wrote two useful functions(one is linear regression and the other one is robust linear regression) to get the relationship between year and lifeExp in gapminder dataset and made some comparisions between these two results. It's good you split the dataset into two parts, and use the fitted model to get the predicted values of the test data.
• Since I didn't choose to work on nested data frame, so I learned a lot from this part of your homework. You are proficient at analyzing data using dplr, join and spread. You give a clear table to compare the different slopes using two different methods. What's more, your multiplot function is really powerful! We could put several plots together using your function and also give each plot a nice header, that's great!

Some improvements:

• Maybe you could choose more exercises in chapter14 and try to learn more about the stringr package.
• Actually, I am a little confused why you chose a density plot to show the relationship between the slope and intercepts in the "challenge" part. I can't figure out the meaning of the plots. If you could explain a bit, I believe it will be better! Overall, a very nice work! Check plus!!

[emwest]

First, you have a really easy to navigate repo. I appreciate how you set up your Readme with links directly to your homework documents.

Your hard work is quite evident in the first part of the assignment. Working with nested dataframes is tricky and having not chosen this option I learned a lot from your homework! I appreciate how you not only developed a nested data frame with gapminder, but you also showed how to back out of one using the unnest() function. You also implemented many of the functions, including graphics, that were highlighted earlier in the semester. While I am not as sure as the other reviewer about the nuances of the statistics, your graphics are very clear and well represented, especially your faceted figures. You optimized clarity in the challenge portion by expressing your output in table form and following with a brief comment.

Having done the stringr exercise myself, I would have liked to seen you pick a few more exercises and maybe a couple of the more challenging options. Based on the earlier part of your homework, your proficiency in this stuff is clear. Overall this is a really strong homework submission! Nice job

[derekcho]

Hi @hannahdxz ! Here are some comments about your hw06:

• Good work for the first task, your functions extended from the examples in several ways. However, I would suggest that you try adding some of the elements that we discussed in class such as error checking. Nice observation at the end about the MSPE and sample size, I was just about to say the same thing. The main point of the task was to write a function and you did it well!
• Your work with a nested data frame was also well done. Each step was well explained and easy to follow. Although not entirely necessary, it might have been useful to show the output for a country with the nested data (for example, show the data and fit for Afghanistan after it has been nested). I like your density plots showing the distribution of life expectancies.
• Very interesting work in task 3! Looking at differences between linear regression and robust regression slopes is a neat way to identify outliers and it seemed like it worked well.
• Well done overall! You performed more than the required number of tasks and each task was well done! Your work was easy to follow because you explained each step well and showed the relevant output. Your coding style is neat and efficient as well and you used effective plots when it was relevant. Nice work!

Note: your mark (check minus, check, check plus) will be distributed later.