Good Data Versus Good Graph?

I saw the original chart above, “Driving is Why You are Fat,” and I liked its graphics and the implicit conclusion within the title. However, I was unsure how the data shown between states correlated to one other or even proved the title. This confusion stemmed from the infographic use of relative icon size & color transparency to measure metrics, so I thought a streamlined simple line chart would clarify things.

mine less pretty but correct: http://goo.gl/HJTLG

In order to clarify if/how state obesity rates correlate to the commuting methods within each state, I sorted the data first by obesity so the states would be ranked from lowest to highest obesity, to create a baseline metric to evaluate how the commuting methods correlated to obesity. I went back to the original data source, combined the pure data into segments that represented the states' working populations' commuting methods, and finally converted all metrics to a percentage of the total state's commuting population.

Working directly with the data led me to question the original motives of the graph since the data doesn’t necessarily map as clearly as the title of the piece leads you to assume. For example, the transportation method metric is based on the total working population (older than 16) of the State, while the obesity metric is the total population (older than 16) of the State.  Obviously, one could discover that the unrepresented non-working population represents the majority of obese people in each State, therefore rendering the transportation method metric of limited value in assessing the health of each state's population.

The outlier in the data is the District of Columbia, which I could remove, since it’s technically closer to a city and not a state, so the population doesn’t have the same commuting requirements. However, I kept in, because it does suggest a new area of data to look into to explore whether or not the implicit hypothesis of the original graph “Driving is Why You are Fat,” which is to segment out the city dwellers in order to see if their commuting methods make them less obese compared to the state average and therefore see if driving to leads to obesity. If you look at the District of Columbia and compare it abstractly to Virginia or Delaware the answer is yes. Unfortunately, I couldn’t find the data to explore this line of questioning more. Perhaps another time.