The original Pop chart and dataset

Redrawing data by using a multiway dot plot

According to the new multiway dot plot, I can easily find scale information of observations  by using table look-up, which is not directly obtain from the original pie chart. For instance, the scale information of the observation with the largest people is (1) major =Commerce (2) Year=”2005″; (3)Student= 720 people.

And viewers can get more accurate physical information from the new dot plot by using Pattern Perception. From the dot plot, I observe that the red symbols as a group are shifted to the right with respect to the blue symbols, which means the number of student at particular major in 2005 is larger than the number of students at that major in 2006. However, this information is impossible to visually get from the pie charts.

Over all, the statement that “Any data that can be encoded by one of these pop charts, such as a pie chart, divided bar chart, can also be decoded by either a dot plot or multiway dot plot that typically provides far more pattern perception and table look-up than the pop-chart encoding” is correct.

The data contain three variables, calories, fat, sugars, from UScereal dataset.

1. Scatterplot matrix

In this Figure the (2,3) panel is a graph of calories on the vertical scale against sugars on the horizontal scale. From the (2,3) panel, I find that the general relationship between calories and sugars is positive. As the sugars increases, the calories also increase.

In this Figure the (3,3) panel is a graph of calories on the vertical scale against fat on the horizontal scale. From the (3,3) panel, I find that the general relationship between calories and fat is positive. As the fat increases, the calories also  increase.

In this Figure the (2,1) panel is a graph of fat on the vertical scale against sugars on the horizontal scale. From the (2,1) panel, I find that the general relationship between sugars and fat is positive. As the fat increases, the sugars also  increase. As the sugars increase, the change of fat become larger.

The two special points marked in red. Based on the original data, one is Grape-Nuts, another one is Great Grains Pecan.

Grape-Nuts: Calories= 440, fat=0, sugars=12

Great Grains Pecan: Calories=363.63, fat=9.09, sugars=12.12

Therefore, Grape-Nuts has zero fat but its calories is higher than others. Great Grains Pecan has much  higher calories but its sugars and fat are not large. These two points do not follow the general pattern.

2. Coplot

This plot condition on Sugars; calories is graphed against fat for six intervals of sugars chosen. Except for panel (2,2), each conditioning on sugars, the dependence of calories on fat has a nonlinear pattern: hockey-stick. On the five panels, the slopes are positive.  the panel(2,2) shows this conditioning on sugars has linear pattern. This suggests that there is no interaction between the two factors; the effect of fat on calories is the same for most values of sugars.

3. Three-dimensional scatterplot

 

From the above plot, I find that the general relationship between calories and sugar is positive. As the sugar increases, the calories also increase.

From the above plot, I find that the general relationship between sugars and fat is positive. As the sugars increases, the fat also  increase. 

There are two special points in the upper space. They do not contain large sugars, but their calories are higher than others.

From the above plot, I find that the general relationship between fat and calories is positive. As the fat increases, the calories also  increase. 

And a special point in the right corner has zero fat and highest calories, which do not follow the general pattern.

Part A:

I plot the mean Capacity as a function of week number, comparing four years 2012-2015.

From the above plot, I find that when the number of week is about 30, the mean of capacity reach a  peak, and the capacity has significant increase from 2012 to 2013, and does not have obvious change for the rest three years.  The another interesting thing is the biggest decreases are from week 35 to the end of year, and reach the minimum on the week 52. In my opinion, the audiences would like spend more time with their families celebrating holiday.

Part B

Form the contour plot using different color set

I select Set1 to be the Palette  in the function Scale_fill_distiller. After running the code, I got the above plot. The reason why I think this plot is better than the plot shown on Canvas is that this contour plot is formed by a set of color which distinguish the different levels clearly. The contour plot shown on Canvas does not give a clear vision.