Avatar of Erik Romijn posted April 12 2013
on Erik Romijn

Open for Business 9: Safety in Bike Like a Local

This is the ninth in a series of posts about Apps for Amsterdam: Open for Business, an initiative by Appsterdam, Amsterdam Economic Board and Waag Society to work with three local start-ups to support them in making successful businesses using open data. I’m participating with Bike Like a Local.

One of the bike rental companies I approached said they didn’t like Bike Like a Local at all, because it can distract tourists and actually make biking less safe for them. Based on a list of research the Ministry of Infrastructure and the Environment provided me with, I looked at possible risks and mitigations in this area.

Distraction risks for cyclists

There are five potential distraction risks involved in using mobile devices in traffic:

  1. physical/motoric distraction caused by operating the device;
  2. visual distraction when looking at the device instead of traffic;
  3. cognitive distraction when the device distracts the user from traffic;
  4. auditive distraction when the device makes sound, and sounds from traffic could be dampened;
  5. effects on moods caused by music or conversation.

The most significant effects are, unsurprisingly, seen in making calls of mobile phones. This is mostly due to the cognitive distraction they case. However, the danger is smaller for cyclists than motorists, because cyclists have more time to evaluate their surroundings. The auditive impact typically plays a larger role with cyclists, because they depend more on audio than motorists. Particularly strong danger is, unsurprisingly, involved in visual distraction, like sending an SMS or reading Facebook.

In around 4% of accidents involving bikes, the cyclist was using a mobile device in some way just before the accident. That could be listening to music, a phone call, typing an SMS, etc. This means 96% of accidents could definitely not have been caused by using a mobile device. Numbers are about the same for listening to music as for being in a phone call, although listening to music is much more common. This suggests the risk of listening to music is much smaller than with phone calls. There also appears to be a discrepancy between perception of risk and actual risk, with the former being significantly higher.

In the debate on whether or not to forbid (handsfree) mobile phone use for motorists, a common argument is that motorists are also allowed to talk to passengers. However, a significant difference is that the passenger is able to adapt to the situation, and change the tempo or complexity. They notice when the driver is busy, or the situation is unclear, and stop talking for a while. With a phone call, the other party does not have such information. This adaption significantly reduces the risk. In addition, low audio quality results in a higher cognitive load. And the passenger can also help reduce the workload of the motorist, by keeping track of road signs, for example.

Regarding music, the type and volume is very significant. For cyclists, the main risk is in reduced auditive attention. High tempo and volume are major factors here. The consensus of all research is that there is the use of mobile devices increase risk a bit, but that there is definitely insufficient basis to forbid their use.

Risks for Bike Like a Local

For Bike Like a Local, the risks are in the cognitive and auditive distractions. There is no interface to interact with, making physical and visual distraction insignificant. And, exploring the city with Bike Like a Local is even less involved than listening to music. Like music, and unlike phone calls, it does not involve the user replying or making decisions, but it’s also not as continuous as music, and will not have a high tempo.

Regarding cognitive distraction, I’m aiming for Bike Like a Local to behave like a car passenger. I want to integrate data about black spots (the most dangerous places in Amsterdam) so that I can warn the user, and make sure the app stays quiet otherwise. This might actually raise their attention level when crossing through a black spot. I’m also researching whether I could reduce the frequency of messages as the noise level increases - possibly indicating busier traffic.

To reduce auditive distraction, I specifically tell users not to use in-ear or closed headphones, and to make sure they can still hear other traffic. If needed, I recommend they only wear one earbud. I’m considering to even force all audio playback to one side only, to enforce the user to only wear one bud. I also want to look at throttling the maximum volume.

In other words, although an app like this can distract the user, research in this field suggests that my features will minimise any distraction. And combined with all the features that improve safety, I’m confident the user will end up saver than without.