Measuring National Low Stress Bicycle Accessibility with OpenStreetMap
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 70 | |
| Author | ||
| License | CC Attribution 3.0 Unported: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/58501 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
|
State of the Map US 201939 / 70
5
7
14
20
26
28
30
31
33
36
38
40
43
44
51
53
55
56
60
61
67
68
69
00:00
Computer animation
Transcript: English(auto-generated)
00:01
Hello everyone, thanks for coming, you know, we're probably weighed down by lunch and snacks But I'll try to keep you awake for the next 20 minutes or so to talk about some bike stuff. My name is Brendan Murphy I'm the lead researcher at a little research group here at the University of Minnesota called the accessibility observatory we're housed within the Center for Transportation Studies and
00:21
We analyze transportation systems and networks transit bike auto PED in terms of how well they get us to places that matter to us our primary metrics are access to jobs but we do access to Healthcare locations or acres of parkland and things like that and we do spatial econometrics analysis to look at how our transportation
00:44
networks change over time and I'm here to talk about a Project that I worked on a lot over the past year of using open street map data for the entire country of the United States to analyze how well those bike networks let us get to job opportunities And also incorporating something called level of traffic stress and I'll unpack all of that
01:05
Over the next 15-20 minutes or so. This is the report that we released this summer It's called access across America biking 2017 There's a lot of stuff that rolls up into this map and this list here and I'll do my best to kind of Unpack things for you and guide you through it. But the end result is that we can say
01:22
Which cities provide the best access by bicycle to job opportunities and of course, New York's number one But Minneapolis here where we are this weekend for state of the map us is number seven. Hmm. That's pretty cool Little bit about the background here We needed to figure out how to measure bike access to destinations
01:45
Another consideration is not all roads are equal and not all routes are equal for bikes. Some are less safe Some are more preferable than others some have paths that are really good to use and people prefer them So we need to have quantified
02:00
ways to waste them to to work with those considerations and Thirdly here. We need a systematic classification system to Classify roadways for how bikeable they are Because we're wanting to do this on a national scale Accessibility is
02:21
Very briefly the ease of reaching valuable destinations and it's a way for us to quantify this So we use shortest path routing from an origin to all possible destinations And you set a time limit say you have 30 minutes you travel on your network you count up where you can get to and then you count up the number of say job opportunities at all of the
02:44
Destinations that you can get to and assign that number to your origin you do that for all possible origins in an area You're interested in and then you get your map and you get your data So we can say things for instance the average person in a city say Minneapolis or Saint Paul might be able to reach
03:01
10,000 jobs in 30 minutes on the bike path network That's just a hypothetical example. And this is what we call a cumulative opportunity metric a Of the methodology we use the tag data in OpenStreetMap, so Ideally, all roadways have tag information about whether or not there is a bicycle facility
03:25
About the number of traffic lanes about the prevailing speed and other geometric properties of the roadway or intersection that give us an idea of how a Cyclist might feel biking on that roadway if they might be comfortable or if they might be a little bit uncomfortable if they're
03:44
dealing with Not having space dedicated to them or dealing with a lot of car traffic Once we have that heuristic figured out we tag roads as one two three or four That's one from the lowest stress possible to four is the highest stress possible and we do this for the entire country
04:03
We then tag intersections based on the stress factors of the roads that connect at the Intersections and this is basically if it's a controlled intersection if there's a traffic light then The intersection is assigned the lower stress factor because you can stop the more stressful road
04:22
Traffic will be stopped and you'll be free to cross when the light is green if it's an uncontrolled intersection We assign the higher factor of all of the connecting roadways Once we do that process we can then calculate our accessibility Metric which again is access to job opportunities at the census block level
04:41
but only using Say the lowest stress ways that we labeled in an area or The one and the two ways or the one and the two and the three and so forth so we get four possible levels of access based on how comfortable you might be as a cyclist and
05:00
Once we do that and we can compare at the city level and between cities Nationally to look at how well bike networks do across the country About the tool chain and data we use census LEHD data, which is the longitudinal employer household dynamics program it gives us information about the
05:21
census block locations of workers in terms of where their jobs are and where their homes are and That lets us calculate the access to the job opportunities and then when we roll it up into weighted Individual figures for a city we weight by the population in each census block
05:41
Using OSM that gives us the bicycle travel network and the ways and nodes that we need for routing For routing purposes. We use open trip planner, which gives us a way to do shortest path routing calculations on a large scale and For parallelization we use the ec2 service via AWS
06:02
When we're running our national calculations every year, we're using a weeks worth of time on 20 ec2 instances at the same time Cores are you know on the order of 64 and 256 gigabytes of memory. So this is a lot of computation
06:21
Why we're using OSM is because it's open source. That's why we're all here. It's great It's community managed and national coverage. This lets us give a pretty uniform baseline to Informing our calculations and allows us to compare city to city Obviously, we all hopefully know about the data quality issues and so forth that we can come into contact with in areas
06:44
where there may be fewer people working on the on OSM, but Since we're primarily looking at cities and comparing urban data, we assume that there's a generally a pretty good level of quality
07:00
There's hopefully in the areas that we're analyzing a robust uniform tagging system and Lastly a nice benefit of using OSM data is that OSM users tend to be pretty engaged and involved around Active transportation and biking and transit access and other connected and interrelated urban issues
07:22
So about the LTS classifications this will give you a little bit better picture. I promise I have pictures and map images here, too the first level is Canonically residential streets. There's very little traffic or protected
07:41
And off-street bicycle facilities those might be the lowest stress and you might be comfortable at biking with your kids on them Number two one step up These are slower streets with mixed traffic. Maybe good on-street bike lanes that don't have physical barriers for protection One more step up is the third level. These are faster streets. These might qualify as
08:04
secondary roads in the primary secondary tertiary Sort of classification system and lastly we have primary roads that might be arterials and they don't have bicycle facilities on them A note about the primary secondary tertiary system is we fall back on that
08:22
classification system of roadways if the tag information for a given OSM way Doesn't tell us anything about the number of lanes or the traffic speed Or other geometry that we need to use to inform our our considerations So I've got some picture examples of the classifications
08:44
This one on top left is actually right over here on Oak Street here on campus at the University of Minnesota, and it's a parking buffered protected bikeway and it runs from our campus a couple blocks down to the river where you can connect to a
09:01
river path Two this is a another example from Minneapolis. This is Park Avenue in South Minneapolis there is a double buffered bike lane on both Park Avenue and Portland Avenue and this is an example where you get a lot of space, but they're not physically protected a
09:23
Shoulder bike lane on an arterial might qualify as LCS 3 since you have space given to you But it's kind of subpar and lastly this is Snelling Avenue Right near the Snelling and University Intersection over this way a bit in st. Paul and you you wouldn't want to buy here a
09:45
Quick note, we're not the only people doing this kind of work. This is something from people for bikes in Boulder, Colorado they have something called the bike network analysis to or analyst tool and They're doing this on a city by city level
10:00
so they also use OSM data and They do very similar classification to us of using the tag information and roadway attributes to Classify roadways as high stress or low stress on the bottom right here. It's blue or red They're only doing low stress or high stress. They're not really break They actually break it down to the four
10:22
But then one of their steps is to group one and two together and three and four together to simplify their classification system And all down the left here. These are different opportunity Opportunity destinations, so they're looking at access to Dentists doctors grocery stores a wider variety of things so their metric is a little bit different
10:43
And they wind up with a meta score at the top left of how How well you can reach a variety of things based on the low stress access Our methodology is is kind of based on some of the work that they do
11:01
So now we get to some maps based on our work this is what you see when you map out the OSM network in Minneapolis and the surrounding area of Only looking at the lowest stress ways that we've labeled as LTS one And then we can add the blue. That's LTS two
11:21
Still some gaps in downtown Minneapolis and around This is what we would consider as the the low stress bike network of the one plus the two If we add three, I hope this isn't too hard to see the orange is the third tier so a little bit higher stress, but still you might be able to bike there if you're brave enough and
11:41
Going all the way up to four the red lines are where I might not even bike But this gives you a complete picture of how we label our roadways here via this kind of classification system And we can also do for other cities this is an example of Washington DC
12:01
they have a lot of a lot of arterials in and out of the downtown core and that's been a challenge for DC DOT to address when they're trying to build out their their low stress bike network So those are the underlying networks These are the access maps that we can create with the accessibility data
12:21
We generate what we're looking at is a census block heat map essentially with each census block Colored by the number of jobs you can reach by biking within 20 minutes and This is for the lowest stress roadways only We can take it up to LTS 2
12:43
So your numbers go up a little bit again LTS 3 your numbers go up a little bit because you're expanding The roadways that you're allowing yourself to travel on And finally to four so we have this we have these data we have these cool maps, but what can we do with it?
13:05
We can compare the two We can compare the access levels between the different types of bike networks that we're looking at And we can say things like if every road felt as safe as an off-street path people could reach 10,000 more or 50% more jobs than those that are reachable on just below stress bike network
13:25
And we can do this on an individual census block level or an entire city level We can talk about what might be considered as an access gap Underserved communities may serve to may stand to gain with more low stress networks being built into them and
13:45
We can use this to spatially identify areas that benefit from new bike infrastructure investments. Oh We call the level four open streets because It's kind of named after the our open streets events
14:02
Here or in the same psychology where you close off streets, and you can walk and bike and there aren't cars there Open the streets accessibility to us is If a street felt as safe like that to bike on or if there weren't cars You could bike everywhere, and that's your level of access at LTS form
14:23
This is a map of that difference that I'm talking about so the lighter colored areas Excuse me the lighter colored areas Show that you can on the bike network reach a larger percent of the jobs that you could reach
14:42
by biking everywhere so that level four is your reference and the access you experience on the bike network is what we're comparing and The lighter color areas you have a higher percentage of access the darker colored areas You have a lower percentage of access, which means maybe your bike network. There isn't as well built out
15:04
Maybe there aren't as many good bike lanes there and so forth So that can help us identify darker colored areas where we need to do better, and this is a data page that This is an example of one of the pages that we have in the report that I mentioned at the very beginning of the talk
15:24
This is for Minneapolis. There's a lot here But I'll just unpack a couple things the first graph here shows us the Access to jobs numbers at every time threshold so we can do this for any time budget and this shows 10 minutes on the left through 20 30 40 all the way up to an hour of travel time and
15:46
The orange line is for the corner called medium stress Which is one plus two plus three streets and the low stress is the one plus the two and the chart on the bottom here
16:01
This is what's one of the things I'm really excited about is that we can compare as I showed in this map we can compare accessibility between the Networks for LTS one two and three to that open streets reference point of where you could get to if you could ride everywhere
16:23
And this circled figure on the right here. That's the percentage of Jobs you can reach on the overall bike network as a percentage of where you could get if you could bike everywhere Minneapolis has a 77.8 percentage here, and that's number one in the country so go Minneapolis
16:41
We've got a we've got a lot more work to do in this kind of area But we're doing not so bad already and we be Portland So Quick note on implications in future work this lets us again ID areas of bike numbers for improvement We can talk about equity
17:00
by looking at which populations of people are disadvantaged most by lack of bike networks We can inform policy and planning by doing things like scenario analysis or analyzing project by project If we know a new bike lane is going to be installed, what does that do to our access levels? We can explore ways to analyze the bike networks directly by looking at things like connectivity
17:26
Identifying islands that are disconnected from the rest of a network and so forth and talk about how to connect different sub networks together to enable people to get more places and lastly we can have this kind of research inform
17:42
Conversations around improving OSM classification of bike facilities So thank you very much. I think I've got maybe a minute or two for questions The question was any thoughts about the last point
18:00
great question Yes, I would love to see greater consistency in the labeling of bike facilities, I think a still open question is how to handle buffered bike lanes and Sharrows and there's I mean, there's a whole range of different solutions that we've implemented within the
18:23
Transportation planning world which is kind of where I come from to help bicyclists get around and to help them way find and we do a pretty good job in OSM of classifying and including that information, but we You know There's so many different solutions out there
18:40
That we need to make sure that we're doing a good job of keeping them all in the same place Cyclopath yeah, I do remember a cyclopath. Yep. Mm-hmm. Any other questions?
19:07
Yeah, I do. Um, if you're curious I can I can send you the github repo The question was about what script I used to assign these values to OSM
19:21
Anyone else? Yes, I'm working on that So I'm in the process of generating a bunch of geo packages that will package up the access data and probably also
19:42
The network data too. I don't personally don't have a You know repo or website built for that right now But if you get in touch with me, I can I can get your hands on some onset data All right. I'm at my 20 minutes. So thank you very much