Here I considered, the world electricity production from all energy sources in 2014. The following pie chart summarizes the dataset. The corresponding dataset can be downloaded from here.
It is clearly evident, from the above pie chart the key is given separately. There are six different categorical variables. That is, grey color represents the Coal category, green color represents the Gas, blue color hydroelectric and so on. However, both table look-up and pattern perception are less efficient for pie charts. The pie charts are drawn with no scales so we need to judge the sizes of angles to infer percentages. This would both slower and less accurate than the scanning and interpolation the graph. Thus I constructed the dot plot below using the above dataset.
According to the above graph, we can clearly perceive the pattern of world electricity production in 2014. However, based on the pie chart, we hardly see any pattern of the world electricity production. So, pattern perception is less efficient for pie charts than for dot plots. Thus, using dot plot we can enhance the pattern perception and table look-up.
Secondly, I consider the stacked bar graph. It is given below. The corresponding dataset is given below. Click here to see the original post.
| Q1 Actual | Q2 Actual | Q1 Budget | Q2 Budget | |
| Coffee | 1900 | 2200 | 2000 | 2000 |
| Tea | 1600 | 1650 | 1500 | 1500 |
| Milk | 1000 | 950 | 1200 | 1200 |
They constructed the following plot using the above dataset.
Based on the above graph, we hardly perceive patterns in the data, because the patterns are not clear. The relationship between Budget and Actual values are not readily apparent. Also, the above graph needs two scans and interpolations which is slower and less accurate. More specifically, the “Q1 Actual” values are graphed by position along a common scale, however the others such that “Q2 Actual”, “Q1 Budget”, and “Q2 Budget” are not visual decoding of these latter three sets of values is less accurate than for the “Q1 Actual” values. To overcome these disadvantages, I constructed the multiway dot plot using the above dataset. It is given below.
It can be clearly seen that the “Q2 Actual”, “Q1 Budget”, and “Q2 Budget”are encoded by position along a common scale. This allows us to perceive patterns, not only “Q1 Actual” but also “Q2 Actual”, “Q1 Budget”, and “Q2 Budget” as well.