In the above plot we are looking at the 118th congress with the house of congress plotted on the x-axis and the age of each member represented by the height of the point. Each point is jittered and has it's alpha value reduced to manage overplotting. Each point is also colored to represent the gender of the representative. This is shown in the legend on the right. On top of the points layer is a box plot to represent the mean age, the quartiles, and the maximum and minimum ages within each house of congress and each gender. By hovering over the plot you can see the actual representative the point represents and their age as well as the metrics attached to the box plot.
In this animation we are looking at data from the gapminder dataset. Each point represents a country. On the x-axis is the life expectancy and on the y-axis is the log of the population. The animations is moving forward in time by year, which can be seen ticking away in the top left of the window. There are a handful of striking motions to note, but the most telling is the point in the African plot that drops to the far left around 1990. This point is Rwanda and the motion coincides with the genocide, the x-value of the point at its lowest point is approximately 29 year. You may also notice that a handful of other points within the African plot have shape declines in life expectancy shortly after Rwanda dips. I would love to hear from anyone who has some insight into why that is. You may also notice a point in the Asian plot that declines nearly as far as Rwanda but does so much slower. This point is representing Cambodia. Like the African plot, the Asian plot also has a handful of troubling points that show large declines in life expectancy.
The data for this plot comes from the “NASAweather” package and we are looking at surface temperature as a function of month of the year. Once again the points layer has been jittered and the alpha value has been lowered in an attempt to control overplotting. In this plot there are two additional layers, a regression layer and a boxplot layer. Each of these layers shows one of their objects for every year in the data. The legend provides the connection between year and color.
This animation is connected to the animation above it. Where the above animation looked at life expectancy and population, this animation focuses on the life expectancy variable. This animation can be used as a supplement to the previous one to connect life expectancy drops with countries, as the country names are listed below each bar. A lower bar represents a point further to the left on the previous animation. As before the year is ticking away in the upper left.
In this interactive plot we are looking at data taken from the World Health Organization. Specifically this is the percent of their GDP that certain countries spend on healthcare from the year 2000 to 2015. Each grouping of bars represents one year and the legend shows which color corresponds to which country. Because of the size of the window this plot is shown in, it is quite hard to distinguish individual bars. The interactivity of the plot helps with this; by hovering over each bar you can see the country represented and the exact percent. Also by clicking on countries in the legend you can remove them from the plot entirely. What is most obvious at a cursory glance is that the United States spends a much larger percent of it’s GDP on healthcare than the other countries shown. For some countries this is not surprising.
This plot is also representing data taken from the World Health Organization’s website. In this case we are looking at per capita spend on healthcare instead of percent of GDP. The year range is the same as above. Visually there appear to be three groups of countries when looking at this plot; the countries that are hovering near zero, the majority that are grouped in the middle, and the three countries that spend the most per capita. Most European countries lie in the middle group while Switzerland, Norway and the United States make up the group at the top. The bottom group consists of Afghanistan, Pakistan, China, India, and Peru. The set of countries is not comprehensive but does cover the high and low ends of healthcare spend internationally. There are a handful of inflection points that could use context as well.
This plot was also taken form the World Health Organization’s website. This time we are looking at immunizations though. This time the year range is from 2008 to 2017 and each set of colored lines and points represent a different immunization. When looking at this plot the legend is very useful for clearing out the mess. Once again, by clicking on immunizations in the legend you can remove that data from all of the plots. There is a lot to take in here, but the first thing that struck me when I saw this was the “W” in Niger’s plot. There was a drop in a couple immunizations in 2011 and 2013 in Niger, but in 2012 the percentage jumped back up to the range where it was before 2011 and after 2013.
Taking a break from health data and looking at something less serious, this plot looks at data of Pokemon from the first 6 generations. Each plot is a generation and each bar represents the proportion of Pokemon in that generation with the main type listed on the x-axis. What I was not expecting to see was the missing bars in each generation. I would have expected to see the first generation plot with the most missing bars, then each subsequent generation fill out more types till all were represented. I would expect this because of the creators making new types over time. Instead each generation is missing at least one type.
In this plot we look at the same Pokemon data from above, this time we are looking at the proportion of Pokemon within a range of Attack values. The size of the color block represents the proportion of pokemon with that attack value that are that main type. For example, the highest attack value pokemon in both generation 4 and 5 have the main type of dragon. Whereas in generation 1 and 6, this title is held by psychic types.
This plot is the same as the plot above, but instead of looking at Attack values, we are now looking at defense values.
This plot was generated from the Congressional Budget Office’s (CBO) 10-year budget predictions. First I scraped the numbers out of their baseline report sheets, cleaned the data, and then compared their predictions to the actuals to get the metric on the x-axis. These plots are faceted by the measure and each density curve is for a different report the CBO has issued. It is important to note when looking at these that both the X and Y axes on each facet plot are different. If this were not the case then most plots would appear featureless.