Improving Google Bicycle Map / Suitability Maps on Google Maps

[Adam West]

Hey everybody,

I'm new to the Bike Geeks and recently joined because I want to be
more active in mapping bicycle routes. I'd like to start a discussion
on two things, but I think they are related. We can split the topics
if there is a desire/need.

1. Most of you have probably noticed Google's recent attempt at
giving bike directions / routes and it's obvious need for
improvement.

I made a comment about getting in touch with Google on how to improve
their map. I know they involved outside groups to help create the
maps, so why not try and reach out? There is a way, see my comment:
http://www.atlantabike.org/googlemapsbikethere

I like where google is going with their map, especially the
highlighting of supposed bike friendly routes (They have a ways to
go). This brings me to my next idea, which I've been thinking about
for a lot longer.

2. I'd like to get the bicycle suitability maps digitized in a
similar way and post them on ABC. Ideally maybe the routes could be
incorporated into A-TRAIN for directions and also be visible there?

See a quick google map I made out of a couple of the suitability maps
to get an idea. I made it to figure out a good way to get to work a
while back.

http://maps.google.com/maps/ms?ie=UTF8&hl=en&msa=0&msid=114490326451434161509.00044610cc61ea933be8a&t=h&z=13&lci=bike

Let me know your thoughts.

Thanks,

Adam

[Jett Marks (Home)]

Hey Adam,

I wanted to take at least a brief look at what you presented before
responding. It was a brief look, but very interesting.

Dave Emory, as you know, and several others have wanted to improve the
quality and availability of the cycling information that "wheels on the
ground" cyclists have picked up by experience. Dave's A-Train is no doubt
one of the best and has gotten well-deserved notice, particularly for its
integration with transit schedules.

I've got a "database" of routes that I've collected over the years. I have
struggled with how to best represent the data so it can be shared with
others and had talked with Dave about representing my data in a format that
A-Train could use. (We haven't gotten too far for pursuit of other
projects.) One interesting feature of my dataset is that it can easily
capture information that other people have entered (from bikely.com, GPX
files, or KML/KMZ files).

Google is interested in collecting our data. I didn't look at what they
want, but they appear to be open to accepting whatever vector format we want
to present them with.

This is where I think I would like to go and why I think it is a good
direction:
- Google's Bike directions would be seen by more people than any site we
would hope to create.
- If done properly, we could make the same data available to both Google and
Dave's A-Train. This could serve as a model for what other cities could do
for their transit systems. (Atlanta deserves to be at the forefront of
something bike related ;).
- Fortunately, early on, cyclists with experience will visit in greater
numbers than non-cyclists. We have a chance to make improvements before
less-experienced cyclists blow off the directions offered by Google. Using
Google for bike directions will catch on as the quality of the data
improves.
- The bike suitability maps are way popular because they are so useful and
accurate. Google Maps would be a platform for 1) expanding those maps to
cover more areas and 2) get this information disseminated to a broader
audience.
- And finally, this is just way cool.

I want to leave you with the app I have assembled for presenting routes:
http://routes.cycling.jettmarks.com/?tag=commute is a map showing actual
commuting routes used by contributing cyclists.

I've also got
- recreational routes from two groups I ride with:
http://routes.cycling.jettmarks.com/?tag=vahivelo and
http://routes.cycling.jettmarks.com/?tag=arches
- hill catalogs: http://routes.cycling.jettmarks.com/?tag=hills
- a survey of cycling facilities within Atlanta city limits:
http://routes.cycling.jettmarks.com/?tag=facility

Like Adam, I'd like to move this discussion forward and engage Google. Who
else is interested?

Jett Marks

Hey everybody,

http://maps.google.com/maps/ms?ie=UTF8&hl=en&msa=0&msid=11449032645143416150
9.00044610cc61ea933be8a&t=h&z=13&lci=bike

Thanks,

Adam

--
You received this message because you are subscribed to the Google Groups
"Atlanta Bike Geeks" group.
To post to this group, send email to atlanta-b...@googlegroups.com.
To unsubscribe from this group, send email to
atlanta-bike-ge...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/atlanta-bike-geeks?hl=en.

[jason...@email.com]

How about a mobile app to collect routes? Overlay enough routes and we should have a pretty good map.

[Jett Marks (Home)]

I like the idea of a mobile app.  What sort of device are you considering?  Some of the routes I've cataloged came from GPS devices that output tracks in GPX format.

 

I can't put my finger (mouse) on it right now, but there was a group that was creating an open source map of Atlanta.  They deployed a small army of people on bikes with GPS devices and got them to systematically capture GPS coordinates for several features.  I'll need to dig that up.

 

Capture is something we definitely want to encourage.  Increasing and promoting the availability of different capture methods will increase the volume of data to base routes upon.  It also help get people invested in making sure the maps are the best they can be.

 

It is interesting that you mention "overlay enough routes" because that's what I noticed when I piled one 50-60% opacity track on top of another.  After a while, the commonly used routes got deeper in color.  This is nice visually.  Once we convert to a vector and node format, we can apply some math to compute the best vectors and then visually check to make sure we're not generating poor recommendations based on those calculations.

 

Jett

[John Tackett]

Was that site http://www.openstreetmap.org/  ??

[Jett Marks (Home)]

Yes John, that is the project (and website) that I was thinking of. 

[Adam West]

Jett,

Thanks for reading and the response. (Sorry for mine being so late...
been on my mind, but haven't gotten to it!)

You say "Google is interested in collecting our data." Did they
contact you or were just saying they're interested in general?

I think your database sounds great! It sounds like your data of
routes is robust. Maybe we could use that data to influence how we
make decisions about suitable roads.

I understand that the bike suitability maps were created by analyzing
the road conditions. I think it would be very interesting to see where
a lot of people bike. But for the suitability maps, I would want to
stick to primarily to analyzing roads based on conditions rather than
based on ridership. There may be roads that are very suitable but you
might not have someone logging it as a route in data. Also, you could
have a lot of people riding roads that are unsafe. I know I biked one
for a while until I was hit (Boulevard NE, lucky for just getting a
broken nose).

I'd like to understand the factors that went into deciding whether a
road was suitable. Do you guys know? Maybe Rebecca?

As far as engaging Dave Emory and Google as you mentioned, great! We
should talk to Google to fix some of the major mistakes in the current
bike map they have (Ponce De Leon). Simultaneously, we could work
with existing data to expand upon the suitability maps and update Dave
and Google as we progress.

Also about these maps...I see some odd straight lines. Do you know
what these are about?

Thanks,

Adam


On Apr 1, 7:53 pm, "Jett Marks \(Home\)" <jettma...@bellsouth.net>
wrote:

> I want to leave you with the app I have assembled for presenting routes:http://routes.cycling.jettmarks.com/?tag=commuteis a map showing actual

> commuting routes used by contributing cyclists.
>
> I've also got

> - recreational routes from two groups I ride with:http://routes.cycling.jettmarks.com/?tag=vahiveloandhttp://routes.cycling.jettmarks.com/?tag=arches

> http://maps.google.com/maps/ms?ie=UTF8&hl=en&msa=0&msid=1144903264514...

[Jett Marks (Home)]

Hey Adam,

I understand how it goes with volunteer work. There's a complex formula
with various inputs and outputs: "effort involved", "fun factor", "coolness
factor", and the term with a rather large coefficient, the "is this really
worth my time" factor. The time equation is complex too: "Job
requirements", "Family requirements", and the one that seems to affect me
more than I think it should: "sleep requirements". I want to encourage and
facilitate as many people contributing as possible so we can make the best
use of any time you have to contribute.

Google has not contacted us, but they do have a site (I still have yet to
explore -- has anyone else?) that explains the format of data they accept.
It also sounds like they would be willing to accommodate an organization
with an existing format that would be worth their while to work on. I
suspect their equation looks different from the volunteer's equation, but
I'm sure some sort of calculation is taking place ;).

The data I've got would probably be useful, but as you've pointed out, there
are limitations.

- The bike suitability map does have criteria (found on the map itself
describing "Rated Roadways"), but a lot of it is subjective. The last time
it was revised, a number of cyclists (more than 30?) gave their opinions
about whether routes or sections of routes should be upgraded/downgraded,
and then we basically recorded the consensus. This appears to work really
well. There is little disagreement about the ratings for various sections
of road.

- The data I've got records routes that cyclists ride. I tend to place
more weight on routes chosen by cyclists whose skills I admire, and also
realize that to connect two really good stretches, you sometimes need to
include a bad stretch too. This is not represented or distinguished in any
way by the routes I've captured, but the suitability map should do so. (On
the other hand, routes are about getting from one point to another while the
suitability map only facilitates this and does not go as far as suggesting a
specific route.)

- Yes, ridership is one thing and objective analysis of specific criteria
is another. The routes I've captured really only indicate we have some data
for those routes that could be analyzed according to the criteria.

Next steps? You make some good suggestions:
- Get in touch with Google to see how we can help improve the maps for
Atlanta.
- Catalog the street assessments and find the gaps we want to fill.

Formalizing the data structure/format would be part of this.

As far as straight line segments, there could be a couple of reasons for
this. I'm glad you're looking at this because it would be good to remove
this confusion.
- I've got a segment or two that I was using for test purposes. It is
possible that I've left one or two of these in the datasets.
- Some of the routes have missing portions in which case the route will
jump from one latitude/longitude pair to another pair a long way away. I
need to patch up or remove those disjointed routes.

If you see something that isn't explained by either of these two reasons,
I'm not sure what it might be. If you have the name of a specific route
(you can click on the route to get the name), I can take a look to see what
is going on. Thanks for checking this out.

Good discussion. I think we've got plenty more we can explore.

Jett

-----Original Message-----
From: atlanta-b...@googlegroups.com
[mailto:atlanta-b...@googlegroups.com]On Behalf Of Adam West
Sent: Friday, April 16, 2010 7:47 PM
To: Atlanta Bike Geeks
Subject: [Atl Bike Geeks] Re: Improving Google Bicycle Map / Suitability
Maps on Google Maps

[Chris Kelly]

I don't have the bandwidth to contribute much at this point, but before everyone goes giving map data to Google, it's probably work a look at the Open Street Map version:

http://www.opencyclemap.org/

They also define a data spec and a process for contributing, and your contributions will be under a much better license than contributing to Google.

[Jett Marks (Home)]

Good points Chris.

 

What Dave Emory was pursuing follows the formats used by the Open Source community.  The tradeoff may be building something freely available to everyone or something that will have the reach that Google has.  The two may not be incompatible.

 

Since we are contributing the data, we can choose which project (if not both) is best suited to take advantage of our knowledge of the Atlanta streets.  We should inform ourselves regarding those options.  I would prefer a licensing agreement that allows us to make the data available to anyone including Google.  We would need to talk to Google to see if they are willing to accept a licensing agreement that leaves our contribution available for use outside of Google.

 

Jett

 

 

-----Original Message-----
From: atlanta-b...@googlegroups.com [mailto:atlanta-b...@googlegroups.com]On Behalf Of Chris Kelly
Sent: Saturday, April 17, 2010 10:33 PM
To: atlanta-b...@googlegroups.com

[Adam West]

Google is listening!

Their bike map no longer has major problem roads designated as very
safe for riding, including Ponce De Leon, Moreland Ave, North Ave and
Boulevard. I for one have reported a few of these. This gives us all
the more reason to contact them in a more formal manner. Of course we
do need to be prepared to cite certain roads, but I don't think that
will be a problem. I'll start looking at roads on the map that I
think shouldn't be "green" routes. I'll look to the bike suitability
map methods to do this in an informal manner on my own. We can then
bring our thoughts to the community (here or elsewhere) to get
consensus.

As far as creating our own data independent of Google, I'm all for it
and will be thinking about the roads I ride on my way to work.

Thoughts?

On Apr 17, 10:49 pm, "Jett Marks \(Home\)" <jettma...@bellsouth.net>

> ...
>
> read more »

[Michael Selik]

I've also had responses from Google when I've complained about a route
taking a very steep hill when a good alternative exists.

Michael