MTA Begins Its First Automated Count of Bicyclists

IMG_2590_1.jpgLook closely and you can see the diamond-shaped loops containing an automated bicycle counter on Fell Street between Scott and Divisadero.

Accurate bicycle and pedestrian counts are too often neglected by traffic engineers but the MTA has begun experimenting with technology that will eventually allow its
engineers to gather the most comprehensive data of bicyclists ever taken in San Francisco.

By the SFBC’s estimate, there are about
120,000 bicyclists in the city, but previous efforts to get official numbers have been half-hearted. MTA interns have conducted manual counts once a year, typically in
the late summer, during peak hours at a few intersections, providing what the SFBC calls a "flimsy snapshot of mode share and usage." The agency also runs manual counts on Bike-to-Work Day.

But earlier this month, the MTA installed a Zelt Inductive Loop, designed by the French company Eco Counter, on Fell Street between Scott and Divisadero. According to Oliver Gajda, MTA’s bicycle program manager, it works like a radio antenna. When metal in a bicycle wheel comes within close proximity of the device it
sends a signal.
The settings are designed to avoid counting cars and motorcycles. It can count several bikes at once and record which direction they’re traveling. According to the vendor’s website:

Each time a bicycle goes over the loop, the system
detects the two wheels’ electromagnetic signature and registers a count.
The loop is 1.5m (5ft) in length and more than one loop may be linked to
the same logger.


One big fault, however, is that it can’t count bicyclists who drift outside the bike lane, or more than an inch away from the sensors. This is a problematic issue for that particular section of Fell Street where the device is installed because bicyclists often move out of the bike lane to avoid the mess of obstructing motorists lined up to get into the gas station on Divisadero. Cheaper gas being offered recently at that gas station isn’t helping.

"It’s something we’re working to address," said Gajda.

IMG_2559.jpgThe glut of cars blocking the bike lane on Fell Street as motorists wait to get into the gas station on Divis.

By comparison, the magnetometers designed by Streetline, which my colleague Matthew Roth wrote about recently in a story about SFPark, can detect metals within five feet of the sensor.  Should Streetline or a vendor with similar technology win the SFPark contract, there could be an opportunity to collect bike count data in the trial areas.

Gajda said the MTA will be working with other vendors to test out a variety of technologies. He said he got excited about the idea after attending a Velo City
where there was a display of bicycle counting technologies
being used in some Nordic countries.

The goal is to purchase 15 automated counters that would be installed
at locations all over the city before the city’s bicycle injunction is
lifted.  Gajda estimates the program will cost about $170,000 and plans to submit
a funding request soon to the San Francisco County Transportation
Authority, which doles out the money for such programs.

"The great thing about looking at this from a purely academic point of
view is that the injunction has kind of put a bubble over San
Francisco. And so people have asked, how much does bicycling increase
if you start to stripe bike lanes? This gives us the opportunity to
have the whole city as a laboratory after the injunction is lifted. If
we can get those counters in before the injunction is lifted then we
can lift that bubble up and start striping bike lanes again and then
we’ll be a model for other cities in terms of before and after

In theory, said Gajda, the data could eventually be beamed to a satellite
where it would allow the MTA to capture it in real time.  It’s all part of a larger effort by the MTA to improve the data it collects as it becomes a more multi-modal agency, said MTA spokesman Judson True.

Andy Thornley, the SFBC program director, is excited at the prospect of having some accurate bicycle data in San Francisco, and hopes all of the automated counters can be installed before Bike to Work Day on May 14th: 

Apart from satisfying academic/professional curiosity, having
good bike counts is essential to casemaking for increased funding and
policy commitment for bike facilities — much bike traffic is
invisible at present, making it difficult to justify allocating money
and street space for better bikeways. By counting bikes well, and
showing the positive effects of investment in bikeways, it’ll be
easier (or anyhow less difficult) to rebalance transportation spending
(and rebalance street space) to do more good things for bike

Photos by Bryan Goebel

  • rzu

    I noticed the loop detectors in the bike lane just the other day. It would be great to have this kind of data for bike lanes all around the city.

  • Xanthe Asher

    This counting system is great and will lead to better bicycle lanes as well as become a tool to measure riders usage which is leverage for increased funding opportunities for bicycle related development projects.

  • This provides an excellent opportunity for warring sides to cook the books. The cyclists can organize group rides that navigate the loop detectors, the anti-cyclists can just double park on top of the loop detector!

  • And in fact the settings should not be designed to “not” detect cars – they should be designed to *differentiate* between cars and bikes – in order to assess the magnitude of the problem of cars driving in and parking in the bike lanes 🙂

  • Jeffrey W. Baker

    If the explanation of the operating principle is correct, this seems like it wouldn’t work. Most bike wheels (and many entire bikes) contain little or no ferrous metals. Can’t these loops detect -any- conductive metal (e.g. aluminum)?

  • @ Jeffrey W. Baker

    Yes, aluminum is fine for detection.

  • Bryan Goebel

    Thanks for the clarification JF that it does pick up aluminum and alloy rims, which I’ve corrected in the story. It does not pick up carbon fiber wheels.

  • Most streets do not have bike lanes.

    Most bike lanes will never have sensors.

    Most bike trips will take place either off the bike lane network or on bike lanes without sensors.

    What’s the math behind taking a specialized subset of a specialized subset and then making broad generalizations from that?


  • CBrinkman

    It will be interesting to know the numbers, even if the counts aren’t perfect the first time, or first few times – three cheers to the Bike Program for starting the process of counting. We already know that Market Street had more cyclists then private autos during the rush hours, it will be great to have numbers for other streets as well. Now – how do we count the missing cyclists? The ones who would cycle if they felt safe?

  • CBrinkman wrote: “how do we count the missing cyclists? The ones who would cycle if they felt safe?”

    We count them once policies are put into place that make it safer to cycle and they do.

    One potential problem with the count is that if the numbers come in low, that gives the anti-bike people like Rob Anderson a toehold to challenge the level of funding for bike facilities. That almost puts the civil rights of cyclists up for a referendum based on usage.

    The question is also one of prioritizing scarce resources. I’d rather see the Bicycle Program tell us how they’re going to proceed at implementing the policies in the Bicycle Plan that promote safety and convenience rather than instrumenting a few bike lanes and implementing some of the more spurious segments in the 1997 bike network.


  • This is the same technology that lets traffic lights detect a bicycle waiting for a green yes?

  • marcos – there are sure to be funding problems for the bike plan and prioritization will occur. A lot of that could come in the form of delaying the pieces that the neighbors whine about. But it’s pretty easy to see which projects have the most oomph.

    My vote for #1 – and it’s not even close. Townsend St.

  • Regular

    I rode by this yesterday at 5 pm. One small problem. The cars waiting to get in to the gas station at Divisadero and Fell are backed up so far that they are blocking this section of the bike lane. So those of us on bikes are swerving around into the adjacent lane, far from the counter. Given that drivers are constantly lining up to get gas here, particularly during rush hour, I suspect that the counter is missing a lot of bike trips.

  • Don Miguel

    I’m an everyday cyclist. I go past that area pretty much 7 days/week, but I generally avoid the Fell St lanes as getting doored here would dump you into 40 mph traffic. I take Hayes instead.

    I’m not looking for encouragement or condemnation on the risk/reward of Fell st lanes 😉 But, I think this type of situation presents an additional dilemma with the data collection.


SFMTA Begins Installing More Bike Counters Around the City

The SFMTA has begun installing 22 automatic bicycle counters at 15 locations throughout the city to help get a more accurate tally of the rising numbers of bicyclists in San Francisco. The Zelt inductive loop counters are placed 1 to 3 inches below the pavement and each time someone pedals over one, the system detects […]

Fell Street Arco Bike Lane Might See Improvements Soon

Click to enlarge. Proposed changes to Fell St. bike lane, soft-hit posts in red. The MTA might have a solution for the dangerous conditions at the Fell Street Arco gas station, where motorists queueing up to buy gas routinely block the bike lane and force cyclists to navigate between bumpers or try their hand a […]