Lately I have taken to trying to clean up the tiger road menace around my area. I found JOSM had some accurate imagery for my area and I even found an updated source for that imagery. I took the time to find public data of the roads and their center lines. I took the time to do GPS tracks around my area and compared all three sources to find them accurate. The road data has both local route numbers, state route numbers as well as full street names. Combining these sources I started manually looking at each road for accuracy, updating them as necessary.
To track these roads I have been putting a source for where the information came from. I then removed the Tiger tags as I have read in many comments is typical. I also tag bridges with their reference numbers and a source for where that came from.
Today I received a barrage of comments on my changesets telling me to stop or I will be reported to the DWG. Telling me that what Iâm doing is disruptive to other mappers, considered a bulk import, impeding others from finding bad data, and should not be tagged with a source except in changesets. This is the first time anyone has commented on any changes Iâve made.
Can I have some clarification on what Iâm doing wrong? What more do I need to do before removing tiger tags?
It doesnât sound like you are doing things wrong to me. Looking through your change sets I see a comment on Changeset: 159793067 | OpenStreetMap
I looked at each of the ways in that change set and donât see anything that upsets me. I am not sure why @DUGA is so upset. Maybe they are using tiger:reviewed=no to track if everything they personally want on a road is tagged. For example, while I like to put surface=* and lanes=* tags on every road I touch, I donât think that is a requirement for removing the TIGER tags.
The changesets are: Changeset: 158114518 | OpenStreetMap - In this case the referenced way I only removed the tiger:review tag. The comment user then updated by removing the remaining tiger tags and fixing a tag I had not added or updated.
Changeset: 159793067 | OpenStreetMap - I verified all information except for the surface. I can guess surfaces based on ariel imagery but I figured it would be better it be done by actual survey. I can start doing that I guess.
Changeset: 158142046 | OpenStreetMap - This was my mistake I guess. I leave the fixme notes for myself and the data I was using at the time was not clear where the driveway went.
Changeset: 158726532 | OpenStreetMap - Accusation of bulk import and telling me to stop sourcing changes individually. Personally I feel that sourcing the changes individually provides more accuracy as Iâm using multiple data sources at once. (Tax parcel maps, geopolitical boundaries, bridge maps, address points, multiple imagery sources) and it would be incorrect to say that the road data comes from all of them, as most are just using to validate positions.
I reviewed the discussion you linked and will wait for a more concrete outcome to the discussion before changing the tags in the future.
I didnât think they would be a requirement as the TIGER data doesnât include that information (Surface/lanes). It does in a way indicate the road type, but I found that not to be accurate in many cases.
There at least one consumer that uses the tiger tags as a rough proxy for surface. A biking app if memory serves.
Another solution is to change tiger:reviewed from no to some other value to indicate the work you have accomplished. I donât think thereâs a current common value that matches well. We can always make up a new value of course. I did a big cleanup a bit ago and the current most common values are:
yes - âI did a full reviewâ, my preference is removal but some folks like this better
aerial or position - I only checked geometry
name - I only checked the name
Another idea is to use the broader paved/unpaved values for surface which are easier to tell from aerials but still implies âsomeone should eventually work out a more specific valueâ. I think that also happens to satisfy the main consumer of the tiger tags for implied surface info.
I also am not a stickler for my own ârulesââthey are just for me. I will usually leave the bare tiger:reviewed=no if I havenât fixed all of those things.
It would be good to have @DUGA here to explain the reasoning.
If you are able to at least add paved or unpaved, that is very helpful. Survey focused applications such as street complete will ask the user if they can refine the surface type beyond simply paved so adding paved/unpaved does not discourage further survey.
I am going to keep my comments sparse because otherwise I will rightfully get in trouble with the moderators, but some contributors in OSM are harder to get on with than others. There is certainly a conversation to be had about what constitutes âreviewedâ, but sometimes people need⊠a little coaxing to enter into conversation.
Simply offering my delight and encouragement to any and all âfixers of TIGER data.â Thank you; OSM loves you. Itâs a big elephant, and as long as we keep eating it one bite at a time, weâll burp someday.
Thanks for that TIGERMap. I was not aware of it and it clearly shows that some roads in areas I frequent that I thought I had totally surveyed I either missed or neglected to deal with the tiger:reviewed tag. I need to go back to those locations and re-survey them.
Upon some discussion we have a disagreement on what âalignedâ means.
When aligning ways I take into consideration the following:
Where the way currently is,
Where the state published public data says the way is,
Where the aerial photos show the way is.
If the difference is less then about 3ft I just leave it be, I donât think any of the available data is accurate enough to dispute it.
Within 6ft I side with the state data as aerial photography needs detailed height data to be corrected and this is a very hilly area. Sometimes buildings appear to be at 45deg angles and Iâm thinking that is probably not right. It is also a few years out of date and new ways may not be shown.
For larger discrepancies I try to use best judgement of the surrounding area, typically siding with the state data, as most of the time it aligns with the current data.
How the commenter is assessing the data, I do not know.
This TIGER data is over a decade old in this area and nobody else has ever reviewed it. Even the commenter has modified many of the ways without ever removing the tags. I am genuinely interested in this community and doing things the best I can, but if what Iâm doing above isnât considered good enough for a review I think Iâm out.
Any review is better than no review, even if you never touch the tiger:reviewed=* key. (I donât anymore, sorry.) That key is a crutch. It shouldâve gone away wholesale back in 2010 when we realized how many different things âreviewâ could mean, but it stuck around. Maybe we wanted to wait until a little more of the country had been reviewed. Maybe we expected all the TIGER tags to go away at the same time, and in the meantime we should just ignore them. Maybe we were afraid of swamping Ito Worldâs OSM Mapper maps of who last touched each street, the ultimate leaderboard for OSM.
Whatever the reason, this essentially deprecated key found a second life as a personal task list for some mappers or a heuristic for cycle.travel to avoid the road. Good for them! But we shouldnât allow relics of the past to become wedges within the community. If differences over this key are dissuading mappers from spending their time on TIGER review or, worse, souring them on mapping, then the key should go away, even at the cost of having to employ more elaborate heuristics. Or, if things are fine in general except for some miscommunication among mappers, hopefully we can work that out without creating more work for everyone.
I also think the changeset comments did not really reflect the community. I find the âsourceâ tag difficult to work with - what if an intersection is reworked and there is a municipal source originally on the road. Do I now have to split the road and delete the source on just that portion? But I donât have any absolute rule against its use.
By the way, hereâs a local tiger reviewed view TIGERMap
I suppose I should explain that I donât see the data source to necessarily be the ultimate authority. I see it as a source that is over a decade more recent than the Tiger data. Thatâs half the reason I was putting them there, so they are easily found for further review. I know there are areas that need to be looked at further, but not because it is Tiger data.
If a survey of the road shows it to be wrong by all means remove the tag, as it is no longer referenced from the source.
Iâm trying to follow the rules of open source and working in the public. Iâm leaving relevant information behind for the next person that comes along, assuming it will be me but knowing it might not.
I guess what Iâm coming to realize is the tiger review tag is being used by others to mean something other than reviewing the data that was imported from Tiger.
Iâm leaving relevant information behind for the next person that comes along, assuming it will be me but knowing it might not.
If you can provide a way for me to track what road you claimed to have reviewed but actually still does not meet the standard that both of us can agree with, like surface details, one way, alignment, road name, maxweight or maxheightâŠI will be fine for you to drop any TIGER tags. Otherwise, leave them as is since I donât know how to figure out after all tags related are removed.
Of course, you have the freedom to do anything as long as you follow the community guidelines, but if your contribution brings difficulties to other people, then we need to talk.
Source tag should always be used in changesets, but it should not be used on points or segments or areas nowadays.
Do I now have to split the road and delete the source on just that portion? But I donât have any absolute rule against its use.
If the purpose is to make the map data complicated, then this sounds like a plan. But honestly, what is the point of doing this other than creating unnecessary data that have no benefits?
At a technical level, this is very straightforward and possibly even doable in JOSM with manageable chunks and rate limits disabled.
I think we would need a clearly stated rationale to remove the tag which satisfies the question:
âDoes the presence of the tiger:reviewed tag cause adverse mapper behavior, or can it just be ignored without editing millions of objects in the US?â
I would also want to hear the viewpoint of @watmildon since he runs TIGERMap.