Which Matters More — A Bike Network’s Connectivity or Its Density?

The "connectivity" of bike infrastructure a city has matters, but not as much as some other aspects. Image: University of Minnesota
Adding to total bike lane mileage without creating a denser network does not seem to affect ridership. Image: University of Minnesota

What’s the secret to designing a bicycle network that will get people riding?

A pair of researchers at the University of Minnesota recently set out to test the theory that a connected bike network — where bike lanes provide continuous routes between many possible destinations — is a major determinant of how many people bike. What they actually found was a little unexpected. Connected bike infrastructure matters, according to the study [PDF], but not as much as the density of bike infrastructure.

UMN’s Jessica Schoner and David Levinson used GIS software to map cities’ bike networks and rank them according to connectivity, size, density, and other factors. (“Connectivity” is basically a measure of the degree to which bike lanes intersect within a city, and “density” is a measure of bike lane mileage within a given area.) Then, using Census data, they determined the relationship between each factor and the number of people who commute by bike.

Bike lane density was the most important factor, with each standard deviation (about 1 kilometer of bike infrastructure per square kilometer) associated with an additional 150 bike commuters per 10,000 commuters. For connectivity, one standard deviation correlated to an additional 37 bike commuters. Other factors — the overall size of the bike network, the directness of routes within the system, and fragmentation (separate clusters of bike lanes within the same city) — were not shown to have a statistically significant effect.

Schoner and Levinson caution that correlation does not equal causation, so it’s unclear whether the dense networks enticed more people to bike, or if higher numbers of cyclists helped create denser bike networks. The study also did not distinguish between protected bike lanes, painted bike lanes, and off-street paths, so it does not account for the degree of separation between cyclists and traffic.

But the study does indicate that the density of bike lanes within a city could be an under-appreciated factor in getting more people to ride. “These findings suggest that cities hoping to maximize the impacts of their bicycle infrastructure investments should first consider densifying their bicycle network before expanding its breadth,” the authors concluded.

←  Older Comments
 →
  • Gezellig

    Actually, large swaths of Portland and Berkeley are really flat or very minimally hilly…and the hilliest parts are where the fewest people live/work. See attached images of topographical maps of those cities at end of message.

    My point with the Shattuck example is if you’re trying to go from one destination on Shattuck to another one half a mile away on flat Shattuck, it not only seems extra silly that the Berkeley Bike Plan expects you to track back to the backstreets but that also to add insult to injury in the area I was referencing there are even some notable inclines on the east-west streets whereas Shattuck is flat.

    While hills will be an issue for some people not riding a bike, it doesn’t explain why a full 93-94% of people in these locales currently aren’t hopping on bikes to get around. The lack of comprehensive low-stress networks is the single-biggest factor.

    From Portland’s own survey:

    Survey after survey and poll after poll has found again and again that the number one reason people do not ride bicycles is because they are afraid to be in the roadway on a bicycle.

    –> The “Strong and the Fearless” comprise perhaps 2,000 or fewer cyclists in Portland, representing fewer than 0.5% of the population. These are the people who will ride in Portland regardless of roadway conditions. They are ‘bicyclists;’ riding is a strong part of their identity and they are generally undeterred by roadway conditions—though likely few are courageous enough to venture too far up West Burnside into the West Hills.

    –> The “Enthused and Confident” are those who have been attracted to cycling in Portland by the significant advances the city has made developing its bikeway network and supporting infrastructure over the past 16 years. They are comfortable sharing the roadway with automotive traffic, but they prefer to do so operating on their own facilities. They are attracted to riding in Portland because there are streets that have been redesigned to make them work well for bicycling. They appreciate bicycle lanes and
    bicycle boulevards.

    –> A much larger demographic, representing the vast majority of Portland’s citizens, are the “interested but concerned.” These residents are curious about bicycling. … They like riding a bicycle…and
    they would like to ride more. But, they are afraid to ride.     They don’t like the cars speeding down their streets. They get nervous thinking about what would happen to them on a bicycle when a driver runs a red light, or guns their cars around them, or passes too closely and too fast. Very few of these people regularly ride bicycles—perhaps 2,000 who will ride through their neighborhoods to the local park or coffee shop, but who will not venture out onto the arterials to the major commercial and employment destinations they frequent. There are probably 300,000 in this group, representing 60% of the city’s
    population. They would ride if they felt safer on the roadways—if cars were slower and less frequent, and if there were more quiet streets with few cars and paths without any cars at all.

    https://www.portlandoregon.gov/transportation/article/237507

    Topo maps of Berkeley and Portland:

  • Good points again. Never rode in Berkeley but was surprised at how relatively stressful riding in Portland was. Nothing that I couldn’t handle since I seem to fit the description of a “Bold and Fearless” (despite the fact that I sometimes ride scared to death!) but could see the main roads being a major barrier. They had me looking for alternate routes.

    I wouldn’t underestimate the slight hilliness of towns deterring the “Curious but Concerned” (“I like the idea of biking and all but you need to sweat up all those hills”). You mention that alternate side street routes can be hillier than the main roads. This was the case in Seattle. The main roads were located where they were because they were the easiest way to traverse the topography. The side streets were super quite and narrow in many neighborhoods but difficult to ride due to the steepness.

    Again, Amsterdam, Copenhagen, even Munich are much flatter. I don’t think its a coincidence that the further north you go in Germany, the flatter the country gets and in general the bicycle mode share goes up. This is where I say the electric assist bike could literally level the playing field.

  • Gezellig

    Yeah, Seattle’s a tough one! That hill between Downtown/Capitol Hill is brutal (though biking from where I was staying in Belltown to Fremont via Westlake wasn’t bad at all).

    That’s something I’m not used to in SF. You may remember this comparison:

    http://groups.ischool.berkeley.edu/mentalmaps/img/barriers.png

    http://www.sustainablecommunitiesindex.org/img/indicators/small/Population_Density.png

    In other words the majority of SF’s core dense/vibrant/public sector/private sector/commercial-filled areas are actually in the flattest swaths of town.

    Of course hills–even slight ones–will always be a deterrent to some people, and more importantly, some trips. After all, the Portland report even points out few of even the Strong & Fearless venture much up the West Hills. But no one really lives/works up there anyway. Presumably one of the reasons cited amongst the 35% No Way No How includes hills, but according to the reports it’s not reason #1.

    I think the point for a city that has a goal of 20% trips by bike is to remember that you don’t need 20% modeshare in, say, Twin Peaks (where as you can see despite being the biggest hill is also one of the least dense areas–as it is comparatively few San Franciscans live in or travel to this area by any mode). You don’t even need 20% of people biking to work.

    The low-hanging fruit are attracting a sizable minority of the ~67% of bike-curious people making all those little trips within 1-3 miles who’d like to hop on a bike more but would prefer it look a bit more like this:

    http://media.oregonlive.com/portland_impact/photo/sespokanewof13th-lkge-a-020510jpg-31ae5548b0be2d9f.jpg

    and this:

    https://a.disquscdn.com/get?url=http%3A%2F%2Fbikeportland.org%2Fwp-content%2Fuploads%2F2013%2F09%2Fintersection-after.png&key=aLcsKF93KUWNciwcdLwL2Q

    Places like Portland have done a decent job with the former, not so much the latter.

  • I like the protected intersection but I feel the safety of protected infra starts to breakdown in hilly environments. I couldn’t imagine wanting to deal with the protected intersection even on a moderate downhill. I think the solutions for our hillier towns and cities will need to go beyond the CROW manual and find new solutions.

  • Gezellig

    Thankfully for places like Portland, Berkeley and SF the greatest need for protected infra is in the core areas–which at least in those cities happen to mostly coincide with the flattest areas (Seattle’s a bit different).

    Btw, don’t forget that despite the stereotypes, the Netherlands is actually no stranger to inclines–especially in its hilly southeastern regions, where grades of 10% and above are not uncommon. Believe it or not, this is the Netherlands:

    http://i114.photobucket.com/albums/n247/hanneketoen/o809/204.jpg

    There are even biking guides for such terrain (“bergop” = “uphill”):

    http://www.reisreporter.be/wp-content/uploads/2010/08/bergop-ndl-730px.jpg

    As David Hembrow points out, even in hilly Limburg, bike modeshare of ~30% is not uncommon:

    http://www.aviewfromthecyclepath.com/2009/10/effect-of-hills-on-cycling.html

    While 30% is still lower than most other places in the Netherlands (and hills may indeed be part of the reason), it shows the kinds of modeshare that can be achieved even in places with hills when the proper infrastructure is put into place.

    Back to the US, when hills do impede core areas, there are already a fair amount of possible strategies to deal with local reality, such as SF’s famous Wiggle:

    This is pretty much *the* way to get from Market to Golden Gate Park/Haight/Sunset/Richmond/etc. with very little incline. Its popularity has only been enhanced by progressive rounds of infrastructure appealing to more than just the Strong & Fearless, as evidenced by bike counts and certainly anecdotal evidence such as that documented in the video above.

    I don’t think there’s much demand for a protected intersection on SF’s Twin Peaks, Portland’s West Hills, or the Berkeley Hills, for example, nor should we really be worrying much about those low-density, low-bike-demand areas.

    I think the clear focus is the low-hanging fruit of building out #minimumgrid spine networks amongst the most vibrant corridors of these kinds of cities that are very flat yet still have at best halfhearted bike infrastructure at this point:

    Financial District, SF, current:
    http://www.sabolc.com/blog/wp-content/san-francisco-market-street-towards-the-ferry-building.jpg

    What could be:

    http://sf.streetsblog.org/wp-content/uploads/sites/3/2011/01/market-rendering1_11-07-01.jpg

    Polk St, currently:

    http://sf.streetsblog.org/2015/03/04/yielding-to-cars-first-merchants-sfmta-board-approves-polk-plan-as-is/

    What could be:

    http://www.connectingthecity.org/wp-content/uploads/2011/01/polk-rendering1-110114B1.jpg

    East Portland:
    http://www.neighborhoodnotes.com/images/posts/0/9/9948_o.jpg

    Land of scenes like this:

    http://bikeportland.org/wp-content/uploads/2013/09/belmont-before.png

    What could be:

    http://bikeportland.org/wp-content/uploads/2013/09/belmont-after.png

    The aforementioned infamous car sewer Shattuck Avenue in Berkeley (still looks pretty much identical to this today, just with newer cars):

    http://www.route40.net/images/postcards/ca-berkeley-shattuck-avenue-C1960.jpg

    What could be:

    https://bikeeastbay.org/sites/default/files/images/Shattuck%20Ave%20South%20Berkeley_Streetmix_800.png

    Etc. etc.

  • Damn man! You type too much! But I love it. Great points all. However, in the hilly SW of the Netherlands, the bike infra I saw (on the internet) looks more like it comes out of the AASHTO guide than the NACTO. Mostly on street bike lanes that cyclists can move in and out of depending on speed. As a more experienced cyclists, this is honestly what I prefer on country roads.

  • Gezellig

    Haha blame it on the long wait I had in the jury duty waiting room 😀

    Yeah, the craziest grades are probably not always the best candidates for protected infra but as with in grades in the NL, the biggest hills in SF tend to be where fewer people live anyway so it’s not really an area to focus on.

    As with in SF, if you turn on both the Terrain and Bicycling layers in Limburg you can see bike routes are prioritized in the flatter areas. However, moderate inclines on roads above certain speed thresholds still seem to trigger protected infra at points. The following grades in the attached screenshots of a street in Limburg and SF are examples of that.

    Though, again, obviously this is not where the majority of people live so it’s not really something we should spend much time worrying about–the low-hanging fruit is building out in the flat, more densely populated areas.

  • Great point again about hilly places not being where most people live but that does breakdown in Germany. German urban planners seem to relish building suburbs on steep slopes and the end result is a preservation of arable land and EVERYBODY has a killer view from their balconies! But this probably explains why E-bikes are about a third of all bike sales in Germany.

  • John Harshbarger

    As much as it takes to be safe. The limit may be 25, but I can guarantee most will do FAR over that. My street is a 25 street, yet car regularly do 50+. You can’t be so narrow minded as to only look at the set limit, but also must consider real world data.

←  Older Comments
 →

ALSO ON STREETSBLOG

Why Aren’t American Bike-Share Systems Living Up to Their Potential?

By Angie Schmitt |
As policy director at the New York City Department of Transportation from 2007 to June, 2014, Jon Orcutt shepherded the nation’s largest bike-share system through the earliest stages of planning, a wide-ranging public engagement process, and, last year, the rollout of hundreds of Citi Bike stations. That makes Orcutt, formerly of Transportation Alternatives and the Tri-State Transportation […]

Bike-Share and Bike Lanes: The Chicken and Egg Debate

By Angie Schmitt |
Which should come first, good bike infrastructure or bike-share? This is a matter of some debate, and we don’t pretend to have a definitive answer. But we do have some instructive research. While it’s obviously ideal to have bike infrastructure prior to embarking on a bike-share program, many communities, like Chattanooga, Tennessee, and Dayton, Ohio, […]

Why Bike Lanes With Lots of Bike Traffic Can Still Appear “Empty”

By Angie Schmitt |
Wherever there is a bike lane, there is probably an angry driver complaining that it is always empty. That tends to be the case even when plenty of people do use the bike lane. And there are reasons for that, writes University of Minnesota professor David Levinson. Mathematical, geometrical reasons. Like the fact that free-flowing bike traffic will look […]

Falling Behind on Protected Bike Lanes? Blame Canada

By Brad Aaron |
“Something big is definitely brewing in Canada.” That’s the word from Michael Andersen at People for Bikes, which monitors Twitter for news on protected bike lanes around the English-speaking world. Vancouver’s investment in bike infrastructure paid off with a 64 percent spike in bike traffic from 2013 to this year. And Andersen says Canada as a whole has […]