For part 1, it is easy to find an interesting visualization and to point out the positive aspects. But finding out its negative aspects and providing solutions need more analytical thinking. Regarding to the five qualities of great visualization, it is easy to make an objective evaluation whether it is truthful, functional, insightful, and enlightening, but beautiful is too subjective to make a compelling argument.
For part 2, the most difficult part for me is to choose the graphical form. I saw some people graphing Likert scale using barchart, with all the bars aligning at the neutral line. This inspired me to recode biden into a categorical variable and show the general trend in a more intuitive way. The programming part is not as difficult as I expected (although heavily relied on Google), and it is really fun!
Great critique, and nice choice of your visualization
Part 2
But the variable is actually measured in a continuous fashion. You have values off of the 10-point interval (generate a histogram with a 1-point binwidth and you'll see what I mean)
Don't hesitate to give your visualization a specific title. "Attitudes toward Joe Biden by gender and party". Okay, but what's the story? What's the takeaway?
Did you evaluate keeping the feeling thermometer as a continuous variable? Then generating a similar graph using boxplots or a violin plot to visualize the distribution of attitudes? I think this still gets your point across without throwing information away. However I also think you make a good defense of your choice here (plus nice skills using data transformation as part of the visualization process).
Good idea changing the color palette. Might I suggest something from the blue color group for cold feelings towards Biden?
Hi Dr. Soltoff,
Below is my reflection on assignment 1:
For part 1, it is easy to find an interesting visualization and to point out the positive aspects. But finding out its negative aspects and providing solutions need more analytical thinking. Regarding to the five qualities of great visualization, it is easy to make an objective evaluation whether it is truthful, functional, insightful, and enlightening, but beautiful is too subjective to make a compelling argument.
For part 2, the most difficult part for me is to choose the graphical form. I saw some people graphing Likert scale using barchart, with all the bars aligning at the neutral line. This inspired me to recode biden into a categorical variable and show the general trend in a more intuitive way. The programming part is not as difficult as I expected (although heavily relied on Google), and it is really fun!