Edinburgh: roadside cycle lane not correctly detected

[osmuser63783]

Hi, not sure if you're happy with feedback from the public here, let me know if not.

I've noticed a mistake in the Edinburgh map you shared on Twitter. There's a cycle lane on Melville Street (https://www.mapillary.com/app/?menu=false&lat=55.9501416&lng=-3.213121275&z=17&pKey=204545522628642&focus=map) that's adjacent to a road, just like the one on Leith Walk, but on the map you shared on Twitter Melville Street is dark (cycle track) and Leith Walk is light (roadside cycle track).

The OSM tagging on both looks correct, so I assume there's something wrong in your classification script. I've added is_sidepath=yes on the Melville Street cycleway to say that it's adjacent to a road, I used the same tag when I drew the Leith Walk one. You could probably use this to help detect a cycleway that's adjacent to a road. Other tags that have been used for the same purpose are cycleway=sidewalk and cycleway=sidepath. But in my experience of OSM, none of these tags are used very widely or consistently.

[Robinlovelace]

Hi @osmuser63783 very happy with this feedback, thank you!

[Robinlovelace]

Cc @joeytalbot could you take a look when you get a chance.

[Robinlovelace]

Initially I thought it was due to issues with Melville Road classification: we have a high bar for off road cycleway: separated cycleways that are near 'major' roads are classified as alongside as per this bit of code (will just look for the link):

[Robinlovelace]

https://github.com/nptscot/osmactive/blob/cd0abc3119592a8d3c189081003bf7adfd835970/R/osmactive.R#L108-L116

[Robinlovelace]

Update: I have a hypothesis about the cause of the missing links related to this bit of code.

https://github.com/nptscot/osmactive/blob/cd0abc3119592a8d3c189081003bf7adfd835970/R/osmactive.R#L129-L140

[Robinlovelace]

In the OSM tag cycleway:both=separate: https://www.openstreetmap.org/way/338672406#map=17/55.95026/-3.20999

However, we omit 'separate' on the following (and subsequent) line:

https://github.com/nptscot/osmactive/blob/cd0abc3119592a8d3c189081003bf7adfd835970/R/osmactive.R#L135

Do you have any R know-how @osmuser63783? You could give it a try if so, happy to talk you through, or wait for me, @joeytalbot, @wangzhao0217 to take a look. Great feedback :pray:

[Robinlovelace]

Heads-up @osmuser63783 I think our classification is correct here, we're including only 'major' roads as 'alongside', and this is not a major road. Should we think about updating the legend, e.g. to 'Alongside Major Road'?

See below, @joeytalbot may have some more comments on this.

[joeytalbot]

Hi, not sure if you're happy with feedback from the public here, let me know if not.

I've noticed a mistake in the Edinburgh map you shared on Twitter. There's a cycle lane on Melville Street (https://www.mapillary.com/app/?menu=false&lat=55.9501416&lng=-3.213121275&z=17&pKey=204545522628642&focus=map) that's adjacent to a road, just like the one on Leith Walk, but on the map you shared on Twitter Melville Street is dark (cycle track) and Leith Walk is light (roadside cycle track).

The OSM tagging on both looks correct, so I assume there's something wrong in your classification script. I've added is_sidepath=yes on the Melville Street cycleway to say that it's adjacent to a road, I used the same tag when I drew the Leith Walk one. You could probably use this to help detect a cycleway that's adjacent to a road. Other tags that have been used for the same purpose are cycleway=sidewalk and cycleway=sidepath. But in my experience of OSM, none of these tags are used very widely or consistently.

Hi @osmuser63783 , thanks for the feedback. We classify these cycle tracks differently because Leith Walk is a primary road, while Melville Street is an unclassified road. Our category "roadside cycle track" relates to cycle tracks along major roads only. We should probably make this definition clearer, so thanks for pointing it out.

[joeytalbot]

Closing for now, please tag us if you have any other feedback.

[osmuser63783]

Yes, @Robinlovelace and @joeytalbot, sorry, I wasn't aware that you were making a distinction between segregated cycle tracks along major roads and those along minor roads. I assumed that both came under roadside cycle track in your three-way classification scheme!