[Talk-us] Road abbreviations

Kam, Kristen kristenk at telenav.com
Fri Aug 15 21:45:04 UTC 2014


Nick,

I have been following the discussion about road name abbreviations and wanted to chime in.

I agree with you that bots miss stuff AND goof up in silly ways. A CLASSIC example of a bot mistake is “Street Joseph Street”. This example has been fixed by a participant in a MapRoulette name fixing challenge. But as you can see below, a bot updated the street name from “St Joseph St” to “Street Joseph Street”.  Example below:

http://www.openstreetmap.org/way/13597392/history

For sure fixing names automatically can save time. But you can only correct so much with the implemented bot logic and sometimes make mistakes (what you’re analyzing might not be what you’re expecting).  As my co-worker has said, humans are the best pattern, spelling error catchers.

It’s pleasing to know you’re taking the time to expand abbreviations and correct other data errors along the way. Keep it up!

Best,

Kristen

---
Kristen Kam
OSM Profile --> http://www.openstreetmap.org/user/KristenK

From: Nick Hocking [mailto:nick.hocking at gmail.com]
Sent: Friday, August 15, 2014 2:14 PM
To: talk-us at openstreetmap.org
Subject: Re: [Talk-us] Road abbreviations

Keith wrote

"Sounds like you fixed this quickly. Would you mind explaining the steps in detail you took to achieve this? Thanks"


I decided to use a set of common abbreviations of
Ct
St
Ave
Dr
Ln
Pl
Pkwy
Cl
Cir
Rd
In Australia I'd propably add "Cct" (Circuit) to the list.
Then for the first "Ct"  which is "Court" I went to overpass-turbo.eu<http://overpass-turbo.eu> and ran the query of

<query type="way">
  <has-kv k="name" regv=" [Cc][Tt]$"/>
  <bbox-query {{bbox}}/><!--this is auto-completed with the
                   current map view coordinates.-->
</query>
<union>
  <item/>
  <recurse type="down"/>
</union>
<print/>
This shows all the ways whose name ends in a space followed by ct (in any capitalization).
With this map in one window and JOSM in another window, I then located the first "Ct" in the Josm window and edited it manually to be "Court"

Often there was a *nest* of abbreviations (St Ln Pl etc...) in the same neighbourhood, so having fixed one way I would hav a quick look around that area. and fix any other abbreviations I found.  Then on to the next "Ct".  I generally worked my way west to east and north to south.

Once I had cleared up all the "Ct" (rerunning the operpass query verifies this) I then started on the "St"  (Streets). About30 or 40 hours later Las Vegas was fixed.

What I will now do I export to CSV all the Las Vegas Road names, then grep out all the good ones  (I.E Court, Street,.... Road) and see what's left.
Any typos I've made while editing (E.G Ctourt) and any ways without suffixes will be left and I can investigate if they can be improved or not (from TIGER 2013).

You'll notice that no bots were run in the making of these fixes. This is for two reasons.  Firstly I'm not a fan of bots since they make assumptions which may not always be true. Secondly, manual editing in JOSM with the TIGER 2013 and OSM data showing, means that I found quite a few missing road names in OSM. Also I found quite a few missing roads and If both TIGER and Bing imagery agreed about the road then I added in the road.
I also found a few typos along the way which I could fix as well.

Ie always found that desk checking against another source (like TIGER) and entire city results in massive improvements. Any city I've mapped I always
desk check against another source and inevitably find lots of typos to fix.

Nick

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20140815/0e94d644/attachment-0001.html>


More information about the Talk-us mailing list