R Tutorial: Adding aesthetics to represent a variable

Key Takeaways

in this lesson we'll review how to represent more than two variables on a single plot using the color symbol and size of glyphs to begin consider this scatter plot of national happiness against life expectancy from the scatter plot we see that happiness scores are positively associated with life expectancy but the plot masks potential information provided by another variable such as income classification in this lesson we'll explore how to represent an additional variable throughout this lesson we'll consider data from the world happiness report which contains information about happiness levels across 141 countries the dataset consists of country level happiness scores where 0 is the lowest and 10 is the highest along with 8 explanatory variables including world region population log GDP per capita the world bank's income classification healthy life expectancy and indices of social support freedom to make life choices generosity and freedom from corruption our goal is to create graphics to help us better understand national happiness one approach to add a third variable to a scatterplot is to change the glyph the geometric object drawn on our plot in canvas to reflect the value of the third variable for example we can change the color of the points to reflect what income classification a country is in to do this we add a mapping in the markers trace specifying that color should reflect the value of income we see that the income classification further helps explain the observed Association another way to encode a categorical variable on a scatterplot is to draw a different symbol for each category to accomplish this we map our categorical variable to the plotting symbol which is what plotly calls a glyph interestingly when we map income to the plotting symbol plotly also uses color to represent the income group this is called double encoding a variable and makes it easier for the reader to perceive differences between the groups color can also be used to represent the values of a quantitative variable for example we can use color to represent the population of a country the resulting graphic is difficult to interpret because the populations of China and India are vastly greater than the other countries making us unable to see any other differences this issue commonly occurs when values of the variable differ by orders of magnitude an easy way to remedy this issue is to take the logarithm of population here we take the logarithm of population using the log attend function within the markers trace after applying the transformation we can see population more clearly though it does not appear to be related to happiness and life expectancy when visualizing a quantitative variable we can also use size to encode the information for example we can map population to the size of a glyph rather than the color unlike color plot Lee does not add a legend to explain what the size represents to polish this graphic we should add hover information about the population and change the axis labels to specify the hover information we add hover info equals the string text and add a text argument to the plot underscore ly canvas wrapping the polished text and the values we wish to add in the pasted function here we add the country name and population to the hover info additionally notice that we use the HTML line break tag so that hover info appears on separate lines to specify more informative X and y axis labels we use the layout function notice that we need to specify each as a list the resulting scatterplot now has labels that are easily understood and his distribution ready let's keep exploring the world happiness data to further review these concepts