Evaluation

[bensoltoff]

Evaluation of final project by XXX

Topic	Excellent	Satisfactory	Needs Work
Coding style		x
Coding strategy		x
Presentation: graphs	x
Presentation: tables	x
Achievement, creativity		x
Ease of access	x

Remarks:

• Very thorough intro to the literature and research question. Excellent start for a seminar paper or master's thesis
• For "Explore Data", you could probably condense the graphs. Rather than separate facets for each province, you could probably just color code each province and draw a line graph since the data is over time. So for the first graph, the code would be

ggplot(data = litigation_china) +
  geom_line(mapping = aes(x = year, y = cases, color = province)) +
  ggtitle("<N of cases comparison>")

• The shiny app would be much more useful if it dynamically generated the graphs (like above on the page), rather than just a printout of the data table
• Make sure you understand exactly how panel linear regression works so you interpret the results correctly - that is beyond this course, but you will want to be confident in your knowledge of the method if you use it in future research. Also keep in mind your number of observations is rather small (only N=57), so I'm not sure how robust your results will be to other model specifications or if they will generalize to all provinces
• How did you determine the economic factor is more important than education? The variables exist on entirely different scales - you cannot just compare the size of the coefficients to make that determination
• As for the functional form of the model, perhaps try plm with polynomial terms, or a generalized additive model (not sure how to estimate that with panel effects)
• Consider integrating your code into the final website - see @JordanAMartin's website for an example of this