Google Street Traffic Inaccuracies
The idea of street level traffic overlays in Google Maps is obviously appealing. In theory, it eliminates the guesswork involved in trip planning during heavy traffic times. Where I live in Los Angeles (universally known for its shallowness congested highways), drivers often abandon the freeways in search of a less congested side street. Knowing the status of that side street before taking your exit should be a huge help.
Google announced this feature, available in major cities, in August. It gathers data from a variety of sources, including pulling speed information from smartphones running Google’s Map application. Herein lies the problem. My experience has shown that this traffic data is less than accurate.
I noticed this problem while driving across town and trying to choose between freeways and side streets showing all yellow and red. I was traveling at night, around 9pm, and suspected that the street traffic may not be as bad as the Maps application on my iPhone indicated. None the less I was worried about this large red patch on my route:
When I arrived to the problem area, I was not surprised to find that the traffic was not nearly as bad as the map was showing. In fact, I snapped this picture right after taking the screenshot of the map:
As you can see, I am the only car on the road! Notice also that traffic on the other side of the street shows yellow on the map, and is also extremely light in reality. I suspect that the problem comes from how Google utilizes the data it gets from mobile users of its Maps application. For example, if the application polls the speed of a user while they are sitting at a red light, and the speed is returned as 0, how does google treat this anomaly? What about when someone is parked in the car and looking up directions? I know that the data displayed by Google is close (but not exactly) in real time, but the red block in question existed for 30 minutes before my arrival and continued to be red for ten minutes after I had passed through. The issue did not seem to be a jam that cleared just before I arrived. Further, traffic jams do not clear immediately, they should change from red to yellow to green as the traffic lightens.
As with any crowd-sourced technology, valuable data comes from the collective average of all users, not from any one specific user. I suspect that the way Google is treating the data relies too heavily on individual results, rather than a more accurate average of many users. These issues are likely to become less of a problem as use of the Maps application becomes more widespread, especially if it is integrated in the iPhone’s Map application.
