Transforming How We Collect and Visualize Information about Urban Accessibility | Hacker News Reader

Transforming How We Collect and Visualize Information about Urban Accessibility

projectsidewalk.io

I have lived without a car for like a decade. It is shocking how much the world is designed for cars. This winds up being a de facto "screw you" to everyone else, including people reliant on public transit. Many people reliant on public transit are either poor or handicapped (thus can't drive) or both.

I wish this project well. Glad to see it on the front page.

As a white middle class male and native english speaker in a western country, deciding to ride my bike to work or the shops was the first time I ever felt like I was in a minority group.

Makes me think of this piece:

https://alittlemoresauce.com/2014/08/20/what-my-bike-has-tau...

(Of course some folks object to that piece: http://grist.org/cities/please-people-stop-using-bikes-to-wh...)

The street I live in the East Bay doesn't even have a sidewalk, or a bike lane for that matter.

Nicely done. This is the kinda of stuff I like to see machine learning and mapping used for.

I have a friend who is in the wheelchair and walking around DC with them it's really apparent how some of the places are just impossible access, or if it is possible it's unnecessarily annoying. Some of the things are fixable such as potholes some are harder, such as the ramp is missing. Some places we just had to give up and change plans because we get to where we were going.

Project results page are throwing a proxy error currently probably because of the HN traffic. In the meantime here is the GH page:

https://github.com/ProjectSidewalk

The idea is commendable, however I find it unfortunate that they did not choose to use already existing data on OpenStreetMap (perhaps because it's a competitor to Google Maps?), for instance reusing the weelchair routing efforts started in 2010: http://wiki.openstreetmap.org/wiki/Wheelchair_routing

OpenStreetMap already has metadata that could not only help with tracing weelchair paths (pedestrian crossings have a "curb" field designed especially for weelchair accessible; POIs such as shops and buildings have a weelchair-accessible boolean; etc.), but also indicate which sidewalks contain, for instance, tactile pavings for the blind, besides lots of other useful metadata.

Also, there is a lot of effort already invested in OpenStreetMap, and more contributors are always welcome. The overall utility of the maps should help ensure the data will remain used by the community at large, and hopefully up-to-date in the long term.

Thanks for your comment. In short, we agree. We have recently been working with OpenSidewalks (https://www.opensidewalks.com/) on figuring out the best way to integrate our data into OSM. We have also been communicating with Wheelmap.org, which is complementary to our efforts as it focuses on POIs rather than sidewalks/streets.

Our progress is slower than we'd like but we are purely an academic project and our team is comprised solely of students and faculty. We are quite resource constrained and so must manage our goals carefully. Plus, we are OSM neophytes and so we are slowly learning about the community and possibilities.

Oh, and though Google provided some initial funding for our project--via a Google Faculty Research Award--they have never, at any point, restricted what we do or aim to do or how we do it.

Thanks again for your feedback. I hope that in 2018 we can focus more of our efforts on OSM integration. If you're connected to that community, would love additional pointers. Thanks.

-Jon Froehlich (PI on Project Sidewalk)

Haha a site about accessibility doesn't support mobile. All I see is a nice sorry~not~for~you page.

It's a site about accessibility in physical world, not about responsive web design.

Would still be nice if it could be used on the go via mobile phone...

We completely agree. This has been on our TODO list since Sept 2016: https://github.com/ProjectSidewalk/SidewalkWebpage/issues/28.... We are working on it!

Note that we are a small, academic project with no professional developers. So, unfortunately, some TODOs take longer than we expect or want.

Thanks for the comment and for the nudge to work even hard on this!

Funded by Google, uses MapBox

I like the idea that universities can take money from corporations and not feel as though they have to bend their work to support their funder's business models. Kudos to Google and the University of Maryland here.

I suppose the level of influence depends on the institution and the degree of corporate support. I’ve been involved in research where PIs were fiercely emphatic about their unwillingness to share findings to funders prior to publication.

That said, this isn’t classic research / journal pub work, and so presumably Google could decide they didn’t like the direction and yank funding or threaten to yank funding.

To be clear, the funding here is via a Google Faculty Research Award, which is a one-time payment largely focused on supporting one grad student for one year.

https://research.google.com/research-outreach.html#/research...

Google is typically fairly hands off with these awards and we have not, at anytime, felt pressure to pursue a certain direction because of Google funding (though of course, there may be implicit biases that we are unaware of). Most of our funding is via the NSF.

Project Sidewalk is quite dependent on Google Streetview and its API. I don't believe there is a good alternative out there currently (Mapillary is a possibility; Cyclomedia is another but they work via private contracts, which would likely be prohibitively expensive for us). So, if Google decides to end Streetview or significantly change its API, then it will dramatically impact Project Sidewalk's viability. Our hope is that Streetview only becomes more robust and updated more frequently, which will make it an even more valuable "urban science" dataset.

Here's a Diamondback article about "Project Sidewalk" from the University of Maryland Human Computer Interaction Lab hackerspace. (HCIL alum here -- go Terps! ;)

http://www.dbknews.com/2016/10/14/university-of-maryland-app...

A UMD team made an app highlighting D.C. areas inaccessible to people with disabilities.

Jon Froehlich, center, the lead researcher for Project Sidewalk, and his two graduate students Soheil Behnezhad, left, and Manaswi Saha, right.

http://hcil.umd.edu/

http://terp.umd.edu/step-inside-a-hackerspace/

What about WheelMap?

Definitely. We've been in contact with them as well, and we hope something comes of it. As I mentioned above, WheelMap focuses on POIs but not how to get to those POIs, so our work is complementary.

Reply on news.ycombinator.com