Dead ends for cars, not for others

[arnt]

Hi,

Sprawlmap classifies some streets as dead ends because they are for cars, even though they aren't for others. Here are two examples:

https://www.sprawlmap.org/#15.9/48.131784/11.615685 is in Munich, Germany. The two red streets near the middle are dead ends only for cars. The upper-right one is connected to the street to its west with a flight of stairs and to the street to its south with a pavement, so cyclists can leave that street in two directions and pedestrians in three. The lower-left one is connected to the street to its east with a path. All of these are in the OSM database. (I'll report another misclassification involving one of these streets in a moment.)

It's not obvious to me that making a city more walkable than drivable constitutes sprawl.

https://www.sprawlmap.org/#15.9/45.061684/11.798268 is in Rovigo, Italy. The six short north-south streets are dead ends for cars, but their northern end joins an east-west path for bicycles and pedestrians in the park. The four east-west streets end near a river, and again bicycles and pedestrians can go along that river. Rovigo seems full of this. https://www.sprawlmap.org/#15.9/45.079906/11.798227 shows seven dead-end streets, all of them actually meet using paths through a pleasant little park.

I have no answer to this, but basing the sprawl index 100% on whether cars can go more than one way, and 0% on bicycles/pedestrians seems wrong.

Arnt

[Adam Millard-Ball]

Thank you for bringing this issue up. We absolutely include pedestrian and bicycle paths in our connectivity measures, and I agree that they are critical to a full measure of connectivity. But we can only do this where they are mapped in OSM.  

I checked and these examples are correctly classified. 

Regarding the two dead ends in https://www.sprawlmap.org/#15.9/48.131784/11.615685:

- the topmost one is connected by a pedestrian path, but not at the end. That's why there is a blue edge (which links to the pedestrian path) and a red edge (which is a dead end):

https://www.openstreetmap.org/search?query=48.131784%2C11.615685#map=19/48.13260/11.61732

- the middle one is technically a deadend in OpenStreetMap, as there is no topological connection to the parallel pedestrian path. This is certainly a limitation of our work, in that it does not resolve these types of limitations in the underlying OSM data.

The same is true of the Italian example; either the pedestrian paths are not mapped in OSM, or there is no connection in OSM between the path and the road.

[cpbl]

By the way, OpenStreetMap of course likes to hear about mistakes:

https://wiki.openstreetmap.org/wiki/Error_reporting

Maybe we should maybe have a comment on sprawlmap.org mentioning that. One 

person has reported that we need to have more prominent acknowledgement to 

OSM on sprawlmap.org. We give them our topmost thanks in publications, but 

have tried to keep the text slim on the web site...

Chris

[arnt]

I do believe that the "about" page could usefully say that the backing data comes from OSM, and perhaps also that it's processed by <link to gitlab repo>.

OSM has excellent coverage of the places where my wife and I live and also the places where we've been on vacation in the past many years ;) More due to my wife than myself, to be fair.

Arnt

[arnt]

As a matter of UI, it might be clearer if you were to show these footpaths, perhaps as narrower, paler lines?

I zoomed closely in using Naviki now (a bicycle navigation app using OSM data) and it agrees with you wrt. Rovigo. Rather suboptimal data there. It seems... reasonable to assume that two streets are connected well enough for pedestrians if they are really close to each other, and there's no sign of buildings between the streets, or of height differences. People will make paths.

Arnt

[cpbl]

Arnt, 

 Thanks for the suggestion. Yes, it will be possible in a near future iteration to revise what we show in the basemap.

In other news, there is now a link under "More" on the sprawlmap site that takes you to the OSM view of whatever your're looking at.  I hope it makes checking the kinds of things you've found easier.

There's also now explicit acknowledgement to OSM and GHSL in the Data box.

Thanks again for the feedback,

Chris