- Mapper since:
- May 07, 2017
One of things I like to do when meeting new mappers is ask them how they discovered and started contributing to OSM. Well my story goes back to the very, very early days of OSM. I heard about it through some forums (I don’t even remember which) and took a look at it. To say that the tools were in their infancy would be an understatement. There were no aerial images to make use of, things were mostly done through GPS tracks. Living in the heart of Silicon Valley, there were already some roads mapped out. There were areas right where I lived that were definitely wrong, but the tools just weren’t there for me to invest time, so I walked away for several years.
Many years later, I was doing some geocaching in my spare time and there was a big stink when geocaching.com moved from Google maps to OSM. At the time, I didn’t get why they would do that. (A while later it became painfully obvious as I realized that Google was charging through the nose for high traffic users of their services.) I didn’t really pick up OSM at that time either. Add another year or two and I saw that Strava had switched from Google to OSM as well and I really started to be more aware that OSM was becoming more and more mainstream. When I retired in 2017, I wanted to have a project to work on so I started learning OSM and finding more and more ways to contribute. The project was so amazing by that point that contributing was simple and the wealth of tools was impressive.
My Projects (in no particular order)
I came across a diary entry about work someone was doing to find all of the sharp angles near on-ramps. Like if you miss the entry for traffic in your direction, there is another entry for opposing traffic that isn’t meant for you. I went through all of the sharp angles I could find (or maybe they generated a dataset) and fixed them all by adding a no-left-turn restriction at that junction.
Pic4Review Bus Stops in Portland
I found a cool project hosted on Pic4Review that was put together by PDX mappers (or maybe even city/transit employees themselves) that put all of the bus stops into P4R and you had to go through each, look at street imagery and tag each stop with the type of bus shelter/seating options that were available. Interesting project and was probably useful to a fair number of people.
OSM Inspector Cleanup
One of the earlier things I discovered was the Inspector put out by Geofabrik and I thought it would be fun to clean out my state of the “duplicate nodes”, “intersecting ways”, and “long ways”. I would click on the map, switch to JOSM, make some changes and save the fix and go find another error to work on. I soon discovered that I could query a map server with a script and get all of the data in a format that I could manipulate and chop up into bits (geohash quadrangles) and then work on all of the errors in an area at once instead of one by one with the map UI.
OSMI Duplicate Nodes
Whenever a way has a list of nodes and two (or more) of those nodes are at the same exact lat/long, OSMI will flag it for fixing. The fix is simple, but the amount of times this error crops up is a little worrisome. I’m not sure why editors aren’t doing better quality assurance checks before submitting updates.
OSMI Self Intersecting Ways
Lots of people make the mistake of closing a loop in a way and then not clicking the final node twice but going back to a previous node and finishing the way there. That will trigger this error getting flagged. There are other similar user errors where mappers will just cross over ways or double back on ways and those all need to be cleaned up.
A strange error that I see a lot is with imports of Microsoft data. For reason, some corners of buildings, instead of going right for example, will go past the corner, and loop back around with a triangle. Of course that’s invalid so I end up cleaning a lot of those up. I’ve tried to find the source of the error so that I could possibly get it cleaned up there instead of having to chase them down later, but I’ve been unsuccessful so far.
OSMI Long Ways
OSM has a limit of 2000 nodes per way. So if you need to draw a detailed river for example, you will likely go over that limit and need to split into multiple ways and/or combine them with multipolygons. However, if you stop at 1999 nodes, you might not notice that this is a “problem”. OSMI sets an arbitrary limit of 1800 nodes. Any way with more nodes gets tagged, and even though it is legal, it does make it a problem for future mappers that might want to clean up the way by adding more points. If you have a less experienced mapper, they might not understand how to split the way or verify that multipolygons that contain that way are modified properly. So I try to make sure that I split those into ways with fewer nodes by splitting and cleaning up as appropriate.
While doing some intersecting ways cleanup with OSMI, I noticed I was coming across unlikely areas that had a boat load of footpaths where no footpaths existed on current aerial imagery. I did some research and found out that PokemonGo players would draw some OSM features around their houses in order to advance in the game unfairly. I wrote a program to dump the database and analyze the ratio of paths to a particular area (in this case, parks). Any park with a high ratio of footpath distance to the area of the park was highly suspect. I was able to cleanup dozens of abusive edits this way. It was also a good project to learn how to import OSM PBF files into a database and do Postgres/PostGIS queries in new and interesting ways.
Combining Similar Ways
While cleaning up intersecting ways, I would notice a lot of footpaths and service ways that were smaller than necessary and adjacent to others with the same (minimal) tags and existed for no apparent reason. So I did some more database imports and PG queries to generate a bunch of ways in my state where I could combine two ways with exact same tags into a single, longer way. It provides no real value other than to make things just a little more tidy.
Golf Course Cleanups
This is a HUGE ongoing project. Just writing about it makes me noxious. Maybe I’ll fill it out later.
HTTPS All The Things
The internet got really popular back in the days before encryption was cheap and a near-necessity. Everyone knows the HTTP acronym, but few of the masses know about HTTPS and its importance in web security. Since secure websites are much more common now and going to an http:// url will usually redirect you to the https:// version, I wanted to make https be more common and visible to everyone. So I’ve written various scripts to comb the entire planet looking for websites that redirect to https and I update the website tag as appropriate. The script takes about a week and then it turns around and starts all over again. I run it under a bot account and the number of changes it makes is quite impressive.