Have you ever traveled out of the state you live in and found yourself saying, “Wow, people in this state are terrible at driving.” Now you can see if your claim was appropriate after looking at the GMAC Insurance National Drivers Test.
GMAC Insurance has been conducting an annual survey where respondents take a driving test that contains questions from DMV tests across the country. Below is a map of the average scores from 2010 along with their inverse ranking among the 50 states and the District of Columbia.
#maker_map_17450 {width: 100%; height: 400px;}
Maker.maker_host=’http://maker.geocommons.com’;Maker.finder_host=’http://finder.geocommons.com’;Maker.core_host=’http://core.geocommons.com’;
Maker.load_map(“maker_map_17450″, “17450″);
From the map you can see that states in the darker orange color range had the highest scores and states with the lighter orange colors scored lower on the test. On the 100 point scoring scale the highest state score was Kansas with 82.3 as their average. The lowest scoring state was New York with a score of 70.0.
There was also a second part to the survey. This part surveyed drivers on the types of distracting behavior that they took part in while driving. These distracting actions include applying makeup, changing clothes, eating, talking on a cell phone, and texting on a cell phone while driving. Below is a map of the percentage of respondents per state that responded to participating in these distracting behaviors.
#maker_map_17457 {width: 100%; height: 400px;}
Maker.maker_host=’http://maker.geocommons.com’;Maker.finder_host=’http://finder.geocommons.com’;Maker.core_host=’http://core.geocommons.com’;
Maker.load_map(“maker_map_17457″, “17457″);
The above data is all very interesting and I wondered to myself what might cause the bad driving statistics? I decided to then correlate the average scores from the GMAC Test with three types of data: 1. Max State Speed Limits by State (to see if fast driving correlated to bad driving) 2. % of Deficient Bridges by State (see if poor road conditions correlate to bad driving) 3. Population Density by State (to see if congestion correlates to bad driving). These are not perfect indicators, but I thought it might be fun to see of any of these numbers might correlate strongy. Below are the maps:
#maker_map_17460 {width: 100%; height: 400px;}
Maker.maker_host=’http://maker.geocommons.com’;Maker.finder_host=’http://finder.geocommons.com’;Maker.core_host=’http://core.geocommons.com’;
Maker.load_map(“maker_map_17460″, “17460″);
The correlations are interesting:

We see that the max speed limit vs. the average scores had a low correlation of .39. So it is probably safe to say that slow max speed limits or high max speed limits do not deter people from being bad or good drivers. Bridge conditions had a slightly stronger correlation at -.49. This is a bit stronger and may hold some weight for arguments sake. Then the last correlation of population density we see as the strongest at -.56. Also not extremely strong but may be something to consider when deciding why people are bad drivers in certain states.
I found the data from GMAC Insurance to be rather interesting and had fun looking at my state and other states that I have traveled through. See what you think of the results and see if you can see why drivers from Kansas score better than drivers from New Jersey.
One Response to Dataset of the Day: Bad Drivers, From State to State
Leave a Reply Cancel reply
About Us
Welcome to the GeoIQ blog. We write about features of our GeoIQ analytics engine, what is new and exciting in the GeoCommons community, and general industry thought leadership and discussions of geospatial data visualization and analysis.
Please explore what we're working on and let us know if you have any questions or ideas!
New GeoCommons Maps- NYJ city barsone
- Israel Outdoors: Where our applicants are from carine
- jets by state cluster barsone
- Maissade Milko5571
- T-Mobile gulyi01
- AOD MODIS gianluca
Recent Comments
- Matt madigan | Istudyweb on Matt Madigan's Beijing Olympic Report: Camels and 100,000 Flower Pots
- Victor on Dataset of the Day: Who is more Generous? Republicans or Democrats?
- Lidya on TechCamp
- Fares on Dataset of the Day: Profitability of the Fortune 1000
- GIS Blogs – GeoBlogs | GIS Lounge on Off the Map Presents Top 25 Blogs in GIS, GeoWeb and Cartography





I love the scatterplots and regressions, that’s a great feature! Mapping the residuals is great, but it would be more meaningful if the residuals were in test score units:
Currently, the residuals tell us, for each state, how the speed limit or population density differs from the average for states with similar test scores.
I think it is more useful to know how the test scores for each state differ from the average for states with similar speed limits or population densities.
You just have the dependent and independent variables on the wrong axes…flip them (test score on the y, independent variable on the x), and the maps of residuals will make more sense.