Issue 1

The Mean Streets of San Francisco

Issue 1 examines data around SF Muni collisions and traffic-related fatalities measured as part of San Francisco’s Vision Zero project.

SF Muni collision data
SFMTA: operators of San Francisco’s much-maligned transportation network. One source of its famously dependable delays that might not have appeared on your radar is the simple fact that Muni vehicles collide with things. A lot.

Muni collects data on collisions by vehicle type in 3 buckets: buses and trollies; light rail vehicles and historic streetcars; and cable cars.

Buses look remarkably bad if we take them at absolute number of collisions per month. But what happens if we adjust for miles driven?

Number of collisionsCollisions per 100k miles
Bus collisions
Light rail collisions
Cable car collisions

Cable cars come out swinging. Granted, they travel drastically fewer miles than their counterparts, but they seem to experience a pretty remarkable and inconsistent number of issues per month. Purely speculation but the ancient technology, open sides, and tourist:non-tourist ridership ratio might have something to do with it.

As far as I can tell SFMTA doesn’t publish incident-level data. However, the aggregate monthly data still paints an interesting picture. Muni has a Fiscal Year 2015-2016 goal of keeping total collisions per 100k miles below 4.1. As long as cable cars remain in circulation it’s looking pretty far-fetched.

Vision Zero data
Speaking of San Francisco setting aggressive goals: there’s a project called Vision Zero that started in 2014 with a goal of reducing traffic-related fatalities in the city (excluding state highways) to zero per year by 2024.

A laudable goal, but it’s not going well. If you follow local news you may have noticed a recent up-tick in car vs pedestrian and cyclist fatalities. In fact for the first half of the year things are going so badly that if they continue at the same rate this will be the worst year for traffic-related fatalities in San Francisco since Vision Zero started.

First half 201620152014

While the number of fatalities are small relative to the city population, the data is pretty shocking. Non-motorists are dying at a dramatically higher rate than drivers in traffic-related incidents. While fault is difficult to discern in a case by case basis, let alone in aggregate, it’s still pretty remarkable that drivers are the least likely to die in traffic.

San Francisco is working diligently to reduce the imbalance (or, hopefully, zero the whole thing out). If you’ve noticed that it’s getting progressively harder to drive more than a few blocks down Market street without getting diverted to a side-street or that Mission street has become much less conducive to car trips it’s because these projects are part of a broader set of infrastructure projects tied to Vision Zero (PDF).

Unfortunately Vision Zero doesn’t publish incident-level data either. However, they do publish an interactive map that can be used to scrape the desired data. Which I did. You can view the cleaned up data on Numeracy. There are interesting fields ripe for exploration like age, location, and incident comments.

Thanks for reading. There’s a ton of interesting San Francisco (and bay area) data out there. I’ll be tackling a new theme every two weeks. Next issue’s theme will be lighter. I’d love to hear from you, so direct your feedback towards @46point87. Follow along on Twitter or get 46.87 Square Miles in your inbox.


46.87 Square Miles is a joint project with Numeracy

46.87 is a regular exploration of San Francisco Bay Area data by Joshua Jenkins in partnership with Numeracy.