In short: All cars manufactured in the USA since 2010 added as a triangle, the size is the CO2 emissions (right-diagonal), Mile-per-gallon (left-diagonal) and Horsepower (upper side). Smaller triangle, “better” car. Colors show whether it is added for city or highway driving. Green is no difference.
Description: If we want to display three dimensions to our data, the natural shape to choose would be a triangle but it is a very unfamiliar shape as it is visually so different from the Cartesian co-ordinates we’re used to. But it can be a great way to show correlations if you get used to what you’re looking at.
There are many different types of car made available since 2000. How could this data be presented to assist the consumer? In this case each car is a triangle, the smaller the triangle, the better the car (given some criteria a consumer could use like CO2 emissions, MPG and Horsepower). This image is the baseline and has all cars on it. Obviously no “ideal” car exists but this image gives an idea of what is possible when choosing a car. Any car a consumer could chose could be overlaid as a single triangle which allows them to compare against what is available. To do this normally would require a number of tables/graphs that just would not be practical for many.
The blue is for highway driving, orange for city and green no difference between the two as the two driving conditions overlap. As horsepower is the same regardless, the horizontal lines are green until they extend out when city driving lowers MPG and increases emissions.
The point at the bottom where MPG meets emissions shows a clear curve that describes how the two are dependent on each other (a … trendline … would show exactly the same thing, but that’s all it would show).
Point of interest: This data is a great example of how the axis and scale is redundant. For example, do you know what 1 horsepower feels like? The unit and number is arbitrary. We have a feeling for what cars are out there from the highest powered sports cars, to the smallest economical car. If we can see both of these displayed then we don’t need the actual numbers, we just relate the car we’re considering to what we are already aware off. Putting a scale on this would be analogous to getting a marker and doing this to a Van Gogh, and then screaming “but it needed a legend!” while being dragged into the back of a police van. Don’t do that.
Technical: Each car manufactured in the USA during 2010-2017 (33,633 in total) is added as a triangle, taken from (https://www.epa.gov/compliance-and-fuel-economy-data/data-cars-used-testing-fuel-economy). The three sides are defined by the CO2 emissions (right-diagonal), mpg (left-diagonal) and horse-power (up). The closer to the center the higher the MPG or HP and the lower the emissions, i.e. the “ideal” car would be a tiny triangle in the middle. The maximum and minimum for each “axis” is defined by the max/min of the data. Each vehicle is added in the red-color channel for city driving and blue for highway driving. The green channel is the similarity between the two summations. As the HP of each car is the same regardless of driving conditions there is only green along the upper side of each triangle.
The Graph Is Dead 