peer evaluation

[kimswchi]

Hi Longxuan and Xiaoran,

Some comments:

• I think the graphs look very good overall, even with minimal color. Simple but elegant.
• Each graph in the first part of the paper tells its intended story very clearly. You can really see some striking trends.
• The addition of the interactive and facet graphs is a nice touch.
• The interactive plot is very neatly done. I especially like how you see two different labels, one for the year and one for the number of new patents, as you hover over the bars.
• The addition of the static faceted graph is considerate. Some people indeed find static images more useful.

Suggestions:

• For the line graphs, the axis labels and other text could be bigger. They’re readable but still strain the eyes a bit. People with poor eyesight might have trouble reading them. Who knows, you might show this paper to an elderly professor and he/she won’t be able to read the text!
• It looks like you used different formatting for the line plots and the bar plots (minus the interactive one). For example, the facet graph has a gray background, while the others have plain white backgrounds. Why not use the same theme for all static graphs?
• I don’t know what’s the best way to do this, but for each graph you could add an indicator to show which years are the dot com bubble so the reader has a focal point. For example, you could add vertical lines representing the beginning and the end of the bubble. Or you could make the bars or part of the line corresponding to the dot com bubble be a different color so they stand out from the rest.
• When you ask, “Are computer patents truly special?” what do you mean by “special”? What’s the significance of the different distributions for each category? It might be obvious to someone well-versed in the literature, but I think it would be better if you explain the question more.
• For the interactive graph, the y-axis rescales every time the category changes. This makes the comparison challenge (“Are computer patents truly special?”) seem kind of deceptive, since the bars always look to be around the same height range but in reality the values could vary by a lot. You could normalize the counts so that the scale is always the same.
• Also for the interactive graph, it’s a little tricky to compare across categories since there’s no baseline and the whole image changes as you move to a different category. One thing you could have done is have a density plot layered on top of the bar plot to show the patent trends for computers. The reader could then compare the computer trend with other trends by comparing the shape of the density line with the bar plot.
• The faceting might not be necessary if you use a density plot with the smoothed trend lines (like a bar chart, but just showing the outline around the tops of the bars), layer the lines for each category on top of each other, and differentiate them by line type or color. It might look a bit messy, though. To make the computer trend stand out you could do something like have only the computer density line be red and the rest be black or gray.

• Soo Wan