After making several iterations of a MapRoulette challenge and a recent diary entry about the topic, I thought it would be a good idea to start an open discussion about the state of SEO spam / brochure-style descriptions in OSM.
My current version of the Overpass query to find spammy descriptions gives more than 3,000 results. Over a thousand of these have already been fixed/removed. This is only in English, and similar queries for other languages would be helpful.
I tried the same query with the note tag, and it shows both fewer results and more false positives, but there is still more than a bit of spam in there.
A related topic that I have not yet explored in depth is logos in the image tag. This could also be worth looking at, e.g. with a query like this one.
Close observers may notice that I exclude results in Ukraine from my queries. This is because there is an ongoing conflict in the country and mapping there is currently discouraged. If you don’t like that you are free to adjust the queries to include this country.
I ran a similar campaign for myself for a while - often just looking for
overly long descriptions yields 95% spam. This is very often accompanied
by a lack of proper tags - often it’s just name, website, description,
and if you’re lucky perhaps an address or phone number. In these cases I
have usually deleted the thing outright because I figured I couldn’t be
bothered to fix whatever the SEO company did wrongly. (Quite frequently
these were in implausible locations as well, such as in the middle of
the road.) If someone made a genuine effort to tag their business and
went over board with the description, I just removed the description.
Regarding notes, there’s an ongoing SEO effort adding businesses with
notes that go “Business Name: XY, Website: XY Keywords: XY…”, and I
delete several of them every day.
I am not a fan of deleting POIs or notes with POI information outright, because they do add value to OSM when they’re properly mapped and tagged. As I mentioned in my diary entry, quite often the spammy descriptions contain useful info for which we have well-defined tagging schemes. If you can’t be bothered to investigate, I recommend leaving it alone for somebody else to fix it with more care at some point in the future. The MR challenges I’m running prove that mappers will do that.
You are right (in that there might be a morsel of information there) but
if I can see that someone apparently dumped this data into OSM in some
kind of “fire and forget” mode (often a throwaway account with just this
one edit, and often something that looks like a paid-for SEO operation)
then I’m not in the mood to spend free volunteer time researching what
the correct tags might have been, nor do I want to encourage anyone else
to do it. It would only send the wrong message - “just throw your shit
at OSM, the community is going to fix it for free”.
The spam edits may be a bit annoying, but they are perfectly valid and legal contributions to the map. With some relatively simple quality assurance by us we can use these edits to actually make the map better, and we certainly don’t need to spend limited OSMF resources on this matter.
In any case “finding a lawyer” as you put it is not a viable solution to this problem, unless you somehow happen to have enough money to pay a lawyer to do our QA tasks for us.
For those who are interested, I just uploaded the fifth iteration of the “Spammy business descriptions” MapRoulette challenge: https://maproulette.org/browse/challenges/41316
I did my best to reduce the number of false positives, and this is the first iteration of the challenge that also features ways.
The Overpass query I used to build the tasks can be found here: overpass turbo