hw07 ready for grading

[hannahdxz]

https://github.com/hannahdxz/STAT545-hw-Dong-Xinzhe/tree/master/hw%2007

a155f80

[yuanjisun]

Hello @hannahdxz,

You meet all requirements of this homework. Good job! Here are some comments.

Pros

1. You finished this homework in two different ways, using both Makefile and R script. Awesome!
2. Makefile works well without any problems. All redundant files are cleaned with “Clear all” and “build all” works fine.
3. Your figures are really nice and they contain a lot of information, much more than just a figure. You also added a title to make it easier to see what the figure is about.
4. All results are summarized clearly in tables, such as the best/worst countries. It is a very efficient way to present your data/results to me. Great.
5. It is very good that you only display the first several rows of the output table, making your report very clean.
6. Very nice dependency diagram in the README file, which helps me understand your work very well. You described your work in detail in README. Great!
7. You used extra packages which are new to me. I learned a lot from your work!
8. You put files related to the report in a separate folder. Great!

Cons/suggestions

1. In the figure of each continent, only the 4 countries are included in the figure. You may try to put all countries in the figure and change the size of the image. Just a suggestion. You already do a great work here.
2. Instead of writing codes for each continent and change the continent names in R scripts, you may want to define your own function. In this case, you can save a lot of time in writing codes and reduce the memory usage of the computer. This will be extremely important if you run a very large dataset. This is just a suggestion.
3. Maybe more explanation should be added to the final report. For example, what is the figure about?
4. It would be better to put your two methods in a separate folder, which can make your repo cleaner.
5. There is no need to render README to HTML. Only keeping MD file is fine.

Overall, I really like your work!

Yuanji

[menglinzhou]

Hi, @hannahdxz !! Great work overall!The code is clean and easy to follow

• All tasks are complete perfectly.
• In your markdown file, you plot very useful graphs, and show what you learned from the class clearly.
• There are no error with your code, all the process work successfully.
• Do your homework in two ways, with both R script and Makefile, which is really brilliant.
• Your readme file is really useful, with a nice dependency diagram. I can clearly know how you process your homework!

Some other notes:

• In 02_statistical_analysis.R , when fit linear regression, maybe you can do it with a functions instead of re-writing the code (the pipeline of filter->arrange->select).
• Also for the part 2, I think it will be better for us to save the sd of the whole model instead of the sd of each parameter estimator, cause it will be more sensible to choose model with the first one. And then, for each country, there would be only one row.
• Because in your Rmd, you make it to be GitHub-document, so maybe in your Makefile, when you render the report, you'd better write 03_gap-report.md: instead of 03_gap-report.html: .

Menglin

[derekcho]

Hi @hannahdxz! Here are some comments about your hw07:

Three or more scripts, an Rmd, and a Makefile: Yes Starts by downloading data, ends with Rmd: Yes The output of each is the modified input of the previous step: Yes Includes some analysis and at least one figure: Yes Makefile includes all scripts and Rmd with correct dependencies: No (see comments) Makefile runs: Yes

• Nice work getting your pipeline to run! I was able to use your Makefile to recreate all of your figures and the Markdown file by just running your Makefile.
• At the end of 01_filter-reorder-plot.R, when you save an RDS file, you should not use the extension “.tsv”. saveRDS saves things using the “.RDS” extension and other programs will not be able to read this “.tsv” file since it really isn’t tab separated. However, your code was still able to run properly.
• For 03_gap-report.html, you list gapminder.tsv and gapminder_reorder.tsv, etc. as dependencies but you don’t really call them in 03_gap-report.Rmd. Instead, you redownload gapminder.tsv and modify it within that file. Although you can do everything in one file, it kind of makes the rest of the pipeline unnecessary. I understand that you want to show each step of your work, which is fine, but you can show R code in Rmarkdown without having it run, and then actually load your dependencies too. Since you produce the plots in previous steps, you can just embed the “.png” file rather than recreating it in the last script.
• Your exploratory analysis was well done. You produced some nice plots and reordered the levels of continent correctly.
• Your statistical analysis was done pretty well for the most part. However, it looks like you mixed the standard errors of both the intercept and slope when you arranged by standard error. You should focus on just one measure standard error (of either intercept or slope), or the variance of the residuals as proposed in the assignment. This is a minor point though since the accuracy of the statistical analysis isn’t really the point of this assignment.
• Nice scatterplots that show some pretty interesting life expectancies.
• I was unable to find a progress report for this assignment
• Overall, your assignment was well done! Aside from a few errors here and there, you were able to get your Makefile to run without errors. It would have been nice to see a few more interpretations of plots here and there though, although your workflow was very easy to follow. My biggest concern is how you basically redid everything from scratch in your R markdown file when it should have just loaded the data and plots from previous sections. Good job!

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