In short: All cars manufactured in the USA since 2000. 9 dimensions in a single image: year, horsepower, engine displacement, emissions levels (Hydrocarbons, CO, CO2, NOx) for city and highway use.
Description: This image is intended to reinvent a scatter-plot (or 2D histogram) in a simple way, but with the aim to include a lot more dimensions. Each car manufactured in the USA during 2000-2017 (60,256 in total) is added as a square. The even geometric nature of the image makes it accessible to look at, though 9 dimensions in the data are displayed. These are available for interpretation (see Technical) and allows the viewer to ignore the aspects of the data that they are not concerned with. It asks the question how many dimensions to the data can a person interpret simultaneously.
For example, to see the improvement in engine design as the engine displacement reduces (smaller squares) for the same horsepower output (along a horizontal line as you go from 2000 at the left to 2017 on the right). At the same time, the range of horsepower engines available widens. An additional aim of good engine management is then the overall reduction in emissions (per mile), and as the colors change from red to yellow the “city” and “highway” emissions balance out. The green then signifies the cars that have higher emissions per mile on the highway than the city, these are the higher power cars (not the greenest cars ironically).
Point of interest: An unexpected aspect of this image was the optical illusion generated from making slight changes to a very geometrical pattern. Though it is an even grid, it looks misaligned and warped. It forces the eye to interpret the data. Is this the same as a trend-line (spit) or does this inadvertently help the data reveal its nature?
Technical: Each car manufactured in the USA during 2000-2017 (60,256 in total) is added as a square, taken from (https://www.epa.gov/compliance-and-fuel-economy-data/data-cars-used-testing-fuel-economy). The position of each square horizontally is the year; vertical is horse-power (with limits set to make the image square). The size of each square is the engine displacement with each square a combination of four tests of engine emission with the darker the color the more the emission per mile (top: Hydrocarbons; bottom: Carbon monoxide; right: Carbon dioxide; left: Nitrogen oxides). The red channel is for highway driving, whereas green is city and blue is set to a constant value when non-zero. The average value of each point is used to normalize for the number of cars produced. Depending on how you define a dimension, 9 dimensions of the data in a single image.
The Graph Is Dead 