Accessibility and Centrality Based Estimation of Urban Pedestrian Activity

Recent working paper

Estimated evening peak period pedestrian activity
Estimated evening peak period pedestrian activity

Non-motorized transportation, particularly including walking and bicycling, are increasingly becoming important modes in modern cities, for reasons including individual and societal wellness, avoiding negative environmental impacts of other modes, and resource availability. Institutions governing development and management of urban areas are increasingly keen to include walking and bicycling in urban planning and engineering; however, proper placement of improvements and treatments depends on the availability of good usage data. This study attempts to predict pedestrian activity at 1123 intersections in the Midwestern, US city of Minneapolis, Minnesota, using scalable and transferable predictive variables such as economic accessibility by sector, betweenness network centrality, and automobile traffic levels. Accessibility to jobs by walking and transit, automobile traffic, and accessibility to certain economic job categories (Education, Finance) were found to be significant predictors of increased pedestrian traffic, while accessibility to other economic job categories (Management, Utilities) were found to be significant predictors of decreased pedestrian traffic. Betweenness centrality was not found to be a significant predictor of pedestrian traffic, however the specific calculation methodology can be further tailored to reflect real-world pedestrian use-cases in urban areas. Accessibility-based analysis may provide city planners and engineers with an additional tool to predict pedestrian and bicycle traffic where counts may be difficult to obtain, or otherwise unavailable.

Begging for Simplicity

Cross-posted from Begging for Simplicity … in which I complain about beg buttons, and general second class treatment for pedestrians.

Begging for Simplicity


Pedestrian actuators call for a pedestrian signal at an intersection which is semi-actuated (where the green-time goes to the mainline except when a vehicle is on the side street, subject to a maximum cycle length and a minimum green time for the side street) or fully-actuated (where the green time is allocated to approaches which are actuated subject to a lot of constraints). When they do this, they also tell the controller to extend the green time (and parallel walk signal) given to a phase to be sufficiently long to allow pedestrians to safely cross.

Replacement Walk Button (Franklin and Seymour)

Decommissioned Walk Button (Franklin and Seymour)

But this has problems. Imagine you are on a side street about to cross a main street and the light turns green for the cars, but the Don’t Walk sign remains (since no Pedestrian actuation was recorded). You did not get to the actuator quickly enough. Should you cross on the Green but against the Don’t Walk, or should you wait almost an entire cycle for this to come around. You may or may not have enough time to make it across.

Second imagine the actuator is broken. It may never give a walk signal. (There aresolutions for that, where the default state of broken is “on” instead of “off”, but that doesn’t seem to be widely deployed). For instance, recently I reported toSeeClickFix a broken pedestrian actuator at Franklin and Seymour Avenue in Southeast Minneapolis, which was corrected within 18 days (i.e. the case was closed within 18 days). I am fairly confident a broken traffic light serving cars would have been corrected sooner.

The next  two “beg buttons” (pedestrian actuators for traffic signals) were recently photographed. The one on the University of Minnesota campus (at Beacon and Harvard) recently had a time exemption added, implying that the actuator need not be pressed between 8 am and 6 pm weekdays. (This time exemption seems to have been removed since the photo was taken). This is an improvement over the previous situation (pushing the button in the middle of the day). Still one expects this will, like so many others, become a placebo button, or just break and make it so there is no pedestrian phase.

Beg button at Harvard and Beacon, University of Minnesota campus

However, the complexity is still needless. Traffic signals on streets with sidewalks (which implies pedestrian traffic either exists or is desired) should ALWAYS have an automatic walk phase, just as every cycle gives green time to cars from every approach. (This is especially true in pedestrian areas like a college campus which has a plan that aims to prioritize walking.) Actuators are fine if they make the walk signal come sooner, but being unpushed should not be used as an excuse not to have a walk phase at all. Car drivers don’t have to go out of their way to press actuators, why should pedestrians?

(If traffic is so low you are concerned the time devoted to a pedestrian phase (~12 seconds (36 ft at 3 fps)?) is too long (will cause too much vehicle delay) for this two lane roadway, maybe it shouldn’t be a signal but instead a stop sign (which requires no pedestrian signal) or a yield sign. This can be implemented with flashing red lights if you must you electrical gear.)

Beg button at Franklin Avenue, East River Road, and 27th Avenue SE

Fortunately it is not as complex as the last, at Minneapolis’s favorite five-way intersection (Franklin Avenue, East River Road and 27th Ave SE), which gives instructions for something that should be tacit. That it is not tacit indicates it is a flawed design. If I can read the instructions, I already know how to cross a street. It is not like pedestrian actuators are a new technology. While I want more information at bus stops, crossing a street should be straight-forward, and not require an 11 line instruction set with five graphics. Sadly there is more information here than at the nearest bus stop.

Update. I found the tweet below, which seems appropriate.

Erik Griswold (@erik_griswold) 11/16/13, 18:13 "Making pedestrians press. buttons to cross streets is the death of the city" -Eric Fischer (@enf) at #transpowest

Open Streets – Minnehaha, and some further comments on the Blue Line and environs |

Cross-posted at Open Streets – Minnehaha, and some further comments on the Blue Line and environs

Open Streets – Minnehaha, and some further comments on the Blue Line and environs


I went with my family to attend Longfellow neighborhood’s Open Streets on Minnehaha on Sunday August 11, 2013. I hadn’t been to an “Open Streets” before, though I had been to various events where streets were closed, and this was like that, but more spacious and far more bikey.


We drove to near the northern entrance, since we don’t really live within walking distance (certainly not for the 3 year old). We entered on the north side, at Lake Street and walked down to 46th Street. From there, we walked along 46th over to the LRT station, which we took back to Lake Street station, whence we returned to the Open Streets festival and onto the car to return home.

Closing streets like this once a year is a great thing. Maybe we should go all in and close all streets in Minneapolis one Sunday just to see what happens. Obviously you cannot get the concentration of people on any street, but it would be an interesting experiment.

This one is clearly driven by the bike folks. The various Bicycle Liberation Fronts were all active. Also ran into a bunch of people I knew, so a disproportionately large percentage of attendees were either transportation professionals or advocates. Also we had an almost impromptu Board Meeting of Streets.MN, but could not quite gather a quorum.

There was a nice demo of the proposed CycleTrack on Minnehaha (And why is this even an issue? Make it a demonstration project and study it to see if it works as well as bike lanes on either side according to various criteria.)

My sense is that the activities were too diffuse for pedestrians. At the speed of bicycles, the distances were appropriate (and I saw lots of bicyclists riding back and forth a couple of times, but the 2.5 miles over which this stretched there was a lot of “dead space” for a street festival.

Why were there so many police? Was someone expecting trouble? This seemed far more than necessary for traffic control. Maybe some were expecting a street fight between the many Mayoral candidates and their tee-shirted hordes?

Also, as my wife bemoaned “where are the food trucks?” I would have expected more for this crowd of people. I know the local restaurants (and the American Legion – which offered an excellent value for drinks and hot dogs) didn’t want or need the competition, but the art festivals had food trucks. Was there some prohibition, were they not invited, or did they just not want to come?

The kids did enjoy the bouncy house and the inexhaustible on-street tap-dancers, as well as some of the musicians and the skateboard demonstrations.

Also there seem to be a lot of yoga studios, massage, acupuncture and alternative medicine places here. I am not sure what it says about Longfellow, but it is a lot like 1970s Northern California in that regard.

So as we reached the end of the trip near 46th, we chose to take the LRT back rather than retrace our steps.

My god, the pedestrian streetscape at the LRT stations is abysmal. And I don’t just mean the lack of integrated street-fronting land use, I mean simple things, like pedestrian crossing signals.

At 46th Street and Hiawatha, we pressed the actuator. While the traffic light turned green, the walk signal did not. This occurred even though we had a protected movement. So I crossed against the pedestrian light, but with the green light, otherwise I would be stuck on the East side of Hiawatha Avenue for a long time. I don’t know what the standard signal pattern is here, and maybe it was screwed up by the LRT vehicle coming from the north forcing the gates down sooner than the engineers thought that pedestrians would be able to completely cross, but it seems it was optimized for cars and against pedestrians.

I had never stopped at this (46th Street) station before. Why are the benches in the areas that are unheated/unshaded, while the heated/shaded area is standing room only? Seems needlessly hostile to patrons, and is unlike at least some of the other stations. Also, why do the electronic message boards not know when the next train will arrive? This is 20 year old technology, developed well before this line was built. Still, far better than typical Minnesota bus stops.

At Lake Street station we exited. I had never had cause to walk around here either. This is worse than the area around 46th.

There is an up, but no down escalator. OK, I have seen this at stations before, it saves money, and my home station Franklin Avenue has no escalator at all. When you exit, and want to go east, first you have to cross freeway exit ramps. Technically Hiawatha Avenue is not a Freeway, and Lake Street is supposed to be an urban street, so I am unclear why this is a SPUI. More importantly, why do I get a pedestrian signal on one ramp, (the west-bound ramp from Hiawatha to Lake) but the other is catch as catch can with a “yield” sign for vehicles exiting Hiawatha going east onto Lake Street? They should at least have a stop sign or signal, given this is an urban environment with a transit stop.

(To be clear, this has all been noted before, see e.g. the Comments on Sam’s post. I have suggested studying grade separation before, and I still think the current situation is just a compromised and dangerous mess.)

At any rate, we did eventually all make it across alive, and walked along an unpleasant-ish segment of Lake back to the festive Open Streets area in front of Gandhi Mahal. There we partook of liquid refreshments and made our exit.

How shoveling and plowing snow now extends the duration of icy sidewalks


Minnesota is getting still more ice/sleet/freezing rain/snow today. We will see piles of snow for a long time. Why will we see piles of snow? Because we piled it up. The snow on top acts as insulation for the snow below.
Imagine there is a winter where we get 1 meter of snow (in fact the average for Minneapolis is 1.26 m of snowfall). On the 1 m wide boulevard between the 1 m wide sidewalk and the street, we get 1.5 meters high of snow, as we transport half the snow from the sidewalk to the boulevard to “clear the sidewalk” (the other half goes into the yard maybe, if we have yards. If we have buildings, it all goes to the boulevard). (I realize this is not all at one time, which complicates the model but does not change the basic point).
If the snow melts and sublimates at 5 cm/day (this varies), (and water (liquid snow) refreezes at night), instead of 20 mornings of icy sidewalks, we have 40 hazardous mornings. We have approximately doubled the problem.
This problem in embiggened when you add the snow plows piling additional snow onto the boulevard. In this case we take a 10 m wide street and pile 5 m x 1 m of snow on the boulevard on each side. Now we have put not just 0.5 m from the sidewalk but an additional 5 m of snow on each Boulevard. So the Boulevard is responsible for storing 5.5 m of snowfall over the course of the winter. The ground absorbs very little since it is frozen.
Fortunately, cars are local urban heat islands, so the road is warmer than the sidewalk, and the snow melts back towards the roads at a faster rate than on the sidewalk. Unfortunately, sidewalks are interrupted by streets, which are now icy and slushy in the gutter in the morning. Similarly, we are fortunate that buildings are urban heat islands. Unfortunately, new buildings are well-insulated, greedily keeping their warmth rather than sharing it with the adjacent sidewalks.
For the purposes of walking, I wonder whether it might be better to not shovel or plow at all, and take the inconvenience of walking (driving) through snow, than to shovel and increase the duration over which we must traverse ice and slush. This might be a case of getting punished for doing the “right thing.”
An alternatives is to actually move the snow off the sidewalk’s upstream water-basin. In large cities, we might take the snow piles and put them to the side away from the walk path. This answers the questions: Surface parking lots, what are they good for?

Sidewalks are Hotting Up

Brendon writes in:

Heating a sidewalk section has climate change implications. I calculate the 26-year cost of your section at $8,722 at the low end and $9,708 at the high end (depending on the discount rate you assign to the future impacts of climate change. I tend to lean towards the higher end). This means your break-even point is 8% to 20% higher, meaning maybe 173 to 192 pedestrians per day. Of course with a carbon tax in place, there would likely be more walkers in some places, meaning heating the sidewalks become feasible in more places.
Now, if you could use waste heat that hasn’t been previously captured to heat sidewalks, as they are proposing to do with the new “interchange” plaza and HERC steam, the carbon footprint becomes effectively zero additional. Much less per kWh/BTU.
Other interesting facts, heating all the sidewalks in Minneapolis with electricity from the grid for one year would produce more greenhouse gases than the disposal of all our solid waste and wastewater does over the same time period. The additional energy consumption would be equal to about 1/3 of the current annual consumption in all residential properties in the city. It would increase the city’s annual electricity consumption by 8%.

He nicely identifies a feedback effect, heating up sidewalks will create more emissions, which will heat the atmosphere, which will eventually negate the need for heating up sidewalks. There must be an equilibrium point here.
More seriously, the use of waste heat is a great idea, especially near the HERC. The problem would be building infrastructure to distribute that more broadly. There might also be waste heat from wastewater (which is still liquid in the winter, and thus warmer than the ground around it) which we don’t capture, or let go to roads, by running sewers under the streets rather than the sidewalks.

Walkable Ice

In the absence of significant global warming, Minnesotans still need to contend with ice on the sidewalks (to be clear, in the presence of significant global warming, we would have other problems; and in the presence of significant global cooling, we would face snow and glaciers rather than freezing rain and ice).
My own house suffers this problem, despite (or because of) snow clearance, ice re-forms on the sidewalks and steps, or freezing rain falls on the cleared sidewalks, making them slick, rather than on snow-covered sidewalks, making them crunchy. Further, water drips from the house and gutters because of ice dams, and then freezes on the ground.
My alma mater, Georgia Tech, while not typically subject to much snow or ice, has many sidewalks just above steam-heat pipes, which would clear those sidewalks pretty readily in most conditions. The University of Minnesota does a pretty good job with snow clearance, all things considered, using a lot of labor and snow clearance machines in the process.
Ice clearance is hard in this freeze-melt cycle, especially when the water has no where to drain because (1) the sidewalks are convex (along either width or length), (2) the boulevards are covered in snow creating no place for run off to go and creating a source for new melted water, (3) the storm drains are covered in snow, and (4) the ground is still frozen and/or the soil above the freeze line is super-saturated.
I see a lot of attention to ice-free roads, and very little for ice-free sidewalks. This would greatly enhance walkability, reduce the likelihood of severe injury, and increase the number of pedestrians.
There are a variety of ways to address icy sidewalks:

  • Mechanical: clearing sidewalks with shovels and pick-axes and snow-bots.
  • Friction: Sand, Grit, Gravel make the ice more walkable (by increasing friction);
  • Chemical: Salt (reduces ice via melting);
  • Radiant: heated sidewalks (using a variety of techniques);
  • Protection: covered sidewalks; and

If we consider the cost of an icy sidewalk equal to the probability of a fall multiplied by the cost of a fall, multiplied by the number of people who face that probability per day, times the number of days the sidewalk is icy, we can get a sense of the amount we should invest to avoid the ice.
Let’s say I fall once a year on the ice (typical), after traveling 2.6 km * 2 times a day * 10 ice days = 52 km. My fall rate: is 1 fall per 52 km of ice.
For a house with 10 m of frontage, with 100 pedestrians a day, it gets 1 km of pedestrian traffic per day. Once every 52 icy days, it will see someone fall.
The cost of a fall is unclear, since most falls are unreported. For reported falls which require medical care, the estimate is on the order of $10,000. Let’s assume 10% of falls require medical attention, meaning the average cost per fall is $1,000.
This implies that every 52 icy days (once every 5.2 years if there are 10 icy days per year), each house with icy sidewalks imposes $1,000 in costs. In that case, if we want to minimize social costs, we should be willing to invest $19 day in effective ice clearance. This is about an hour of labor (or two hours of undergraduate labor) to operate simple machines plus some cheap (Friction or Chemical based) treatments). Unfortunately, I am unclear whether $19/day is effective.
We could add delay costs, due to people walking slower on ice, which I estimate to be about a 10% reduction in walking speed. With a travel speed typically of 1.44 m/s, we might decrease that to 1.3 m/s. So instead of the 100 pedestrians taking 7 seconds each to walk in front of the house, they are taking 7.7 seconds. That is 70 person-seconds per day, which has an economic value of (@ $15/hour) of $0.30 per day, two orders of magnitude lower than the fall costs, and so not really worth discussing further.
But can we prevent the ice from forming?
For $1000 every 5.2 years, we get $5000 for a 26 year expected life of a capital investment. If we can make a capital investment of less than $5000 to eliminate falls on our public sidewalk, it would be socially worthwhile.
The cost of heating sidewalks is about $20 per square foot (or about $215 per square meter). A 10 meter by 2 meter sidewalk is 20 meters square, giving us a cost of $4305.
We must consider operating costs, which are estimated at $.60/hour. If it is operating 240 hours per year (this is a guess, I don’t know how long it needs to operate to keep the sidewalk ice free), this is $144 year. (You might run it to melt snow, but that has fewer benefits, just avoiding shoveling, not reduced falling in this simple model, so I don’t consider that). $144 per year is $3744 over 26 years (no discounting), so is a large fraction of the capital costs.
Unfortunately, $4305+$3744 > $5000, so 100 pedestrians is not enough to justify heating. However 160 pedestrians would be a break-even point.
Covering the sidewalks (200m of roofing) could cost $80/square foot ($860/square meter). This lasts 15 years. For 20 square meters, this costs $17,200, well out of range for our residential sidewalk if the only objective is ice reduction, especially since it only lasts 15 years. It might have other benefits, such as reducing our exposure to nature and street-life though.
Policy recommendation: Use student labor to clear sidewalks with low pedestrian flows. Heat sidewalks which have high pedestrian flows. Cover sidewalks with very high pedestrian flows.
Yes, I did fall this year. This post was written between my vertical and horizontal positions, so I apologize in advance for its rushed nature.

Stairing us in the face





There are lots of social engineering messages to get people to take the stairs instead of the
elevator. I normally do this if I can, and agree that it is probably healthier (though the energy  savings is small).

However, the state of our stairs is disrepair and disgrace. Many of our buildings were not designed for this new trend, and stairs seemed to be only intended for fire emergencies.
To wit, some pictures of my favorite staircases, which I use regularly at the University of Minnesota. Clearly no one has got the message that stairs should be at least as attractive as elevators, even if they are fire emergency stairs.The first photo is a staircase at the Washington Avenue Parking Ramp (Garage for those outside the Midwest). It is at least painted, and has windows, but one would hardly call it nice. It functions not just as a transportation corridor for people, but also for drains, making it easy to service, like your utility room.
The second photo is from the same building, but a different staircase. Not even as attractive as the first. Without windows or natural light, not carpeted nor tiled, the walls painted with an undifferentiated institutional color.

The third photo is from the Civil Engineering building, this is a side entrance, not intended by the architects as anything but for service, yet it is the fastest way in and out of that highly circuitous building from the south and it gets a lot of traffic.

Compare this with your most recent elevator ride. If it was the CE building, it was admittedly equally decrepit, but the elevators there are under repair. In other buildings, the elevator is usually a much a nicer ride. Why?

If we want people to take the stairs, let’s make the stairs just a little bit nicer.

QRious sidewalks


ST sends me to Rio (via AP) which reports Bar codes on sidewalks give tourist info:

“Rio de Janeiro is mixing technology with tradition to provide tourists information about the city by embedding bar codes into the black and white mosaic sidewalks that are a symbol of the city.”

This might be a solution to improving navigability, though I think it will puzzle archeologists in 1000 years. The problem of course is it makes people look (1) at their phones rather than the city, and (2) at the sidewalk instead of what’s in front of them.

Walk Minneapolis

The Twin Cities should have something like Walk London (only better). I don’t want just trails (I am familiar with the Grand Rounds, but there should be more), but actual urban paths I might want to take because they are walkable, interesting, and minimize conflicts with traffic. These paths should not simply be on a website or mobile app, but either be marked or signed, or otherwise self-navigating.

In London you have:

The route is indicated on the ground by a variety of signs and waymarks, which are very similar to those of the London Loop. In open spaces they consist mostly of a simple white disc, mounted on wooden posts and containing a directional arrow with the Big Ben logo in blue and text in green (but note that in Richmond black replaces green due to local conservation area considerations). A word of warning: the arrow’s direction may not be clear until you are close up. It is easy to assume that it points ahead, but it may turn – look closely before continuing.
On streets the posts are replaced by larger aluminium signs strapped to lampposts and other street furniture, and additionally carry a walking man symbol. On link routes to stations the word ‘link’ is incorporated into the logo. At major focal points you will also meet tall green and white signposts that give distances to three points in either direction. Some of these locations may also have the big, round-topped information boards.

And of course, they should be contiguous.
The best I can find is this, which helps me if I am a planner, but not a pedestrian. At Bike Walk Twin Cities, which feels like , let’s be honest, Bike Bike Twin Cities, the “maps” link has links to 8 different bike maps on their maps page, only one of which is really only for hiking too, and that is for outstate. The Walking maps page leads me to the useless City of Minneapolis page, the route planner from Metro Transit, and two Skyway maps.
Maybe there is some other resource I am missing. Maybe someone has a grant to do this. Maybe someone had a grant to do this, but didn’t do it.