The most recent project in my data visualization class had students designing small multiple visualizations. According to Edward Tufte, in his book The Visual Display of Quantitative Information:
Well-designed small multiples are
- inevitably comparative
- deftly multivariate
- shrunken, high-density graphics
- usually based on a large data matrix
- drawn almost entirely with data-ink
- efficient in interpretation
- often narrative in content, showing shifts in the relationship between variables as the index variable changes (thereby revealing interaction or multiplicative effects).
Tatsu Johnson successfully met these criteria in his chart visualizing highway safety.
Obviously, you can’t make much out of that. First I’ll give you his description at the top of the chart. Then I’ll show you a few details.
This chart compares the highway safety of all 50 states in the US. Fatalities, drivers and population, as well as the ratio between fatalities/population, are represented as a percentage of the most dangerous states. The lower the percentage of the fatalities/drivers and fatalities/population ratio, the safer the states are in comparison to other states. The ones with the higher percentages are the most dangerous to drive in. SOURCE: Google public data, U.S. Highway Statistics.
Tatsu has employed star charts to represent each state. Every spoke represents one of the variables.
Looking at California, you can see that there are a high number of fatalities, but this is to be expected, given that the state has a large population and a lot of drivers. The most important values in understanding safety, as Tatsu explained, are the two ratios, which are average.
Other points of interest he points out are highlighted in red and green.
Connecticut and Massachusetts rank among the most dangerous states in the US with the highest percentage of fatalities per driver/population.
Mississippi, North Dakota and Wyoming are the safest states, with the best fatality per driver/population ratio.
I guess Tatsu felt that including the star template behind each state provided a better measure. I would argue that they are unnecessary, and in fact, make it more difficult to see the data shapes. The borders are also unnecessary, but were a convenient way to bring attention to states of particular note. These are minor issues in an otherwise beautifully executed visualization.