We probably all think the number of people using their phones while driving is high, but never had the data to prove it. Well, with the results of Zendrive's recent study of 3 million drivers, now we do! But before we get into analyzing that data, here's an interesting photo of a rare double rainbow that a friend of mine took with their phone ... yes, while driving (don't worry J.M. -- I won't out you!)
And now, let's analyze some data! ... The Zendrive study used smartphone apps to detect phone use while driving (ie, "when the driver handles the phone for a certain period of time for various purposes such as talking, texting or navigating"). Their three month study involved 3 million drivers, 570 million trips, and 5.6 billion miles of driving. Their main finding was that drivers are using their phones during 88% of their trips -- wow!
They also provided some more detailed data at the state level, showing the average percent of time people spent using their phones while driving. Here's a portion of their table:
It's great to show the data in a table, but I was glad to see they had also plotted it on a map!
Their map was 'interesting,' and I could quickly identify the states with the highest values. But it was difficult to match the multitude of gradient shades from the map to the legend. I decided to create a SAS version of the map using only 5 gradient shades of color, to see if that might help. I think it's a nice improvement!
But then I started coming up with other ideas that might make the map even better ... for example, using quantile binning. In my next version of the map, each color is assigned to about 20% of the states (quintiles). Whereas the previous map made it easy to identify just the highest two states, this map makes it easier to identify the highest (and lowest) 20% of the states. I think this map is generally more useful, but looking at both maps is, of course, even better than just looking at the data in one way.
There's one other detail the original map had, that my map doesn't ... they included a footnote listing the states that ban hand-held devices. Their footnote was quite long, and it was cumbersome to read all the states, and try to mentally locate them on the map, and keep track of them all. I decided to represent the information visually, by annotating a no symbol on each of those states.
And now for the pop quiz ...
- What factors cause certain states to have lower (or higher) phone usage while driving?
- Since data were collected via smartphone apps, does it undercount people with flip phones?
- What's your opinion on using phones while driving?
- Where do you fit into these numbers? (higher, or lower?)
Feel free to discuss this topic in the comments section!
This blog was reposted from the SAS Learning Post. Go there to view the original.