hw07 ready for grading

[CassKon]

Here is my homework 7!

[qiaoyuet]

Peer Review:

Report:

• For the sorting data part, I'm not sure whether arrange by country the year would implement the new continent ordering. I think if you arrange by life expectancy then it makes more sense.
• Good thing to know that ggplot can take stat_summary() as an add-on, although I think it is similar if you do geom_boxplot().
• I would suggest adding titles to plots would help understanding.
• Good to use saveRDS to preserve the revised ordering of levels.
• ddply seems to be a very convenient function. It would be nicer if you call summaryStats and show the output in the report.
• I found it difficult to understand how you just assign the "best" to "Top" then it would become "worst". It would help a lot if you show the output dataframe.
• For the fitted line plots, I think it would be more consistent if you choose method = "lm" so it matches what you do from the above.
• Especially for plots in continent Africa, I would suggest filtering some countries in the previous steps and showing less countries for a continent, otherwise the plot is really difficult to read.

Scripts and makefile:

• I like that you put all the scripts file links in the readme. Really saves time for me.
• I think putting gapminderData<-read.delim("gapminder.tsv") in the 01_filter_reorder script is more intuitive than putting it in the download-data script, just to make things more coherent.
• Maybe you can try the other version of make file that uses make as a tool to build everything. I think your current one can only run all the scripts together but it cannot render a rmd on itself?

[bcahn7]

Hello @CassKon , Here are some comments on your hw07. Hope it helps!

• You totally understood the homework topic, how to automate data-analysis pipeline.

• In your 01_filter_reorder.R file, you could use geom_boxplot() instead of stat_summary(). geom_boxplot() shows min, 1q, median, 3q, max (it does not show mean value). I guess you wanted to show the mean values because you reordered continent on the mean values of lifeExp and might have wanted to compare this with mean values of gdpPercap.

• I like the way you analyzed by using life expectancy scaled by population. Summary statistics, some descriptions, or title would supplement the plots.

• In your 02_aggregate.R file, I didn't know the function ddply(). It was good to know! You combined this with a customized function to print out all the regression results for each country.

• It seems like your Best, Worst, and Top arranged by slope value are all same. I think you probably forgot to delete Top and desc()

• You did a good job with making customized functions!

• I think it would be better if you could facet by less number of countries or just by continent.

• Your Makefile.R totally works well automating a data-analytical pipeline. It would be good if you could try an alternative way using Makefile to run the pipeline.

• It would be a little bit better if you could include Rmd file we usually use for assignments and see how it works and renders RMarkdown report.

Overall, I think you understood how to automate the pipeline and provided some analysis. You did a good job! Hope this comment helps.