I was curious if the sensor would pick up other things like trees or other cyclist, but it seems like they accounted for that:

> We then log a sensor events [sic] if the majority of cells in the sensor frame agree to the same value within a threshold parameter [...]. This ensures that sensor events are only logged when large objects like cars block the sensor’s field-of-view , i.e., one or more small objects like branches or distance pedestrians in the sensor’s field-of-view will not trigger this condition. While there is no guarantee that this approach strictly identifies cars, we empirically saw during testing that passing cyclists and pedestrians rarely satisfied this condition at the typical passing distance due to the wide field-of-view of the VL53L8.

Also interesting that it's quite cheap to build:

> The whole system can cost less than $25 [...]

From the paper https://dl.acm.org/doi/10.1145/3706598.3713325

In my area there has been a program for years where you can sign up to mount a device like this to your bike for pretty much this exact same purpose. From memory it goes behind your seatpost thougho which seems less annoying.

Needs Strava integration

As a long-time cyclist and former bike courier, I think most of the proximity concerns are probably of my own doing. I wonder if the device somehow accounts for this.

My initial reaction is that an accelerometer might be a better data-point, or combining this with accelerometer data.

I'm working on the assumption that a smoother path means I am interacting less with traffic or other hazards.

Just to plug a friend… Velo.ai does similar things… but has other stuff going on: https://www.velo.ai/

Interesting to see how these two would compare, but my first (light) glance points to velo.ai being further along…

I don’t know Seattle so I’d be curious to know if the proximity and accident hotspots are also high traffic zones in general, whether they have a bike like (and how it’s placed), and if the routes are even bike routes or just routes that riders comfortable jostling in traffic like me took. Comfortable riders may also skew the data by being willing to “lane split” at red lights to pass stopped cars rather than waiting at the back in lane.

Having biked a lot in SF, my impression is the best protected bike lanes are on wide roads like Folsom/Howard, Fell/Oak, etc. where proximity isn’t generally an issue, but I’d expect intersections to be riskier due to higher car speeds. While cars passing on isn’t an issue on the Wiggle with a critical mass of riders, on neighborhood streets where sharing the road is obligated the drivers can be scariest, especially in the Sunset. In NYC, an abundance of one lane, one way streets make controlling an entire street easier.

The reality of city design at the moment is almost any bike route will require the sharing the road with cars at some point, usually at the start and end of a ride, because bike lane and “bike route” coverage is often poor in residential areas and business districts.