Suspected undocumented automated edit of Tesla Supercharger names

Communities United States
1

I first saw this change when updating a different tag on a Tesla Supercharger station in the US. The change goes against US custom (and what the NSI and therefore iD currently recommends), which is name=Tesla Supercharger. So along with my other change, I reverted the name. I saw some more, reverted some more, then realized it was going to be an edit war where whoever saves last wins. I searched and didn’t find it, so I’m reporting it as an undocumented, automated, and unauthorized edit by @stillhart. Maybe it’s buried somewhere.

Each edit was done as an individual changeset, but all have the changeset comment “update charger”, no changeset source, and no imagery. Borrowing from Hitchens’ razor (What can be asserted without evidence can also be dismissed without evidence.), what is changed without a source can also be reverted without a source.

The changesets all have a single tag: created_by=MinimalRubyOSM

If it looks like an automated edit, walks like an automated edit, and quacks like an automated edit, it’s probably an automated edit. You can see where @stillhart was thinking about doing this last year.

So what happened? Let’s take a look by comparing 2 stations:

Station 1) Tesla Supercharger at 909 Story Rd, San Jose, CA 95122
——————————————————————————————
Tesla names this station the same on their website, in Tesla vehicles, in non-Tesla vehicles that access the Tesla Supercharger network API, and in the Tesla smartphone app as:

Tesla Supercharger
San Jose, CA - Story Road

On OSM it’s mapped as a node.

Pre-name change the relevant tags were:

branch=San Jose, CA - Story Road
brand=Tesla Supercharger
name=Tesla Supercharger
operator=Tesla

post-name change:

branch=San Jose, CA - Story Road
brand=Tesla Supercharger
name=San Jose Supercharger
operator=Tesla

Aside from adding ref:tesla=27319, which unlike ref:supercharge_info=* is computable from website=https://www.tesla.com/findus/location/supercharger/27319 and therefore redundant, the only tag change is to name=*:

name=Tesla Supercharger > name=San Jose Supercharger

Station 2) Tesla Supercharger at 1179 S De Anza Blvd, San Jose, CA 95129
——————————————————————————————
Tesla names this station as:

Tesla Supercharger
San Jose, CA - South De Anza Boulevard

On OSM it’s also mapped as a node.

pre-name change the relevant tags were:

branch=San Jose, CA - South De Anza Boulevard
brand=Tesla Supercharger
name=Tesla Supercharger
operator=Tesla

post-name change:

branch=San Jose, CA - South De Anza Boulevard
brand=Tesla Supercharger
name=San Jose Supercharger
operator=Tesla

Again, aside from adding ref:tesla=35274, the only tag change is to name=*:

name=Tesla Supercharger > name=San Jose Supercharger

——————————————————————————————
So now we have all the Tesla Supercharger stations in San Jose, CA with the same name, but a slightly different one that nobody charging in San Jose has ever seen. Is this an improvement? You already know the answer.

Where did this renaming idea come from? On an active carto GitHub discussion about what tag to display, Norwegian mapper Gazer75 said,

We have been very deliberate in the name= value in Norway which contains both the brand and site name from the CPO. Makes it easy to search for without requiring the search engine to also check for brand and operator.

So, if stillhart was following Norwegian practice, the rename would have been:

name=Tesla San Jose Supercharger
or maybe:
name=Tesla San Jose, CA Supercharger
or because there are 21 Superchargers in San Jose, maybe:
name=Tesla San Jose, CA - Story Road Supercharger

That’s a bit messy, but it would hack up the name for Carto in San Jose similarly to how the name is hacked up for Carto in Norway. Maybe that was the intent, but the execution was sloppy? Who knows, but in the US there’s consensus that name=* and brand=* are usually identical, and the NSI uses name=Tesla Supercharger.

I did a check on Canada, Mexico, France, and the UK, and the stillhart edits happened to those countries as well, with variations. I would have done a global check but I kept getting server timeouts.

So, what is going on here, and was this an undocumented, unauthorized automated edit? I searched and didn’t see any discussion about it, but maybe I missed it.

1 Like
2

I posted in Changeset: 180601155 | OpenStreetMap

Almost certainly it was undocumented, unauthorized automated edit violating Automated Edits code of conduct - OpenStreetMap Wiki and therefore contacting Data Working Group - OpenStreetMap Wiki and asking them to mass-revert it would be a good next step.

Though maybe somehow it is explainable so maybe giving them some time (until they make next edit or typical time between their edits so far, whichever smaller) would be OK.

But it is blatant enough that contacting DWG immediately also is likely fine.

3

Thanks, I was thinking about first contacting the DWG, but then I thought maybe it was documented somewhere but not well publicized, which is exactly what happened with the infamous bulk import from supercharge.info last 2025-06-22, which we are still cleaning up after because the importer refused to clean up and left the scene (though still an OSM user and still making edits to that data despite not doing any cleaning of the bad bits that still exist).

4

Hi, writing this on the go, so keeping it short. I wrote my own conflator/editor/map and went through each supercharger by hand but to speed things up by focusing on the „standard“ tagging of supercharger. Looking at the timing would reveal it’s not an import. You can literally see my breaks in between. There are many tags and there were tons of mistakes and mistagging which I wanted to clean up (including that dual import you mention) and mapper confusion with destination chargers. It’s a changeset by supercharger since I literally looked at every supercharger (multiple times). Convention according to my knowledge is that name can be „[location] supercharger“, same same as branch named of locals banks (even discussed as such in a changeset), same as it was in the now deleted wiki and present in a lot of superchargers. I wrote my own tag statistics analyzer to just discover that. (And all the weirdly tagged supercharges to have a common set of attributes )

5

Btw should be mostly clean now to my understanding.

6

As much as I know iD specifically allows for „[location|Tesla] Supercharger“

7

Regarding naming practices. Austria completely disallows names on chargers since they are not nearly never visible on site.

8

I would support reverting, for example I would not set the name of Way: ‪Taco John's‬ (‪237830653‬) | OpenStreetMap to “Spearfish Taco John’s” and yet the supercharger is “Spearfish Supercharger” (Node: ‪Spearfish Supercharger‬ (‪12944743623‬) | OpenStreetMap). “Tesla Supercharger” is more appropriate and fits with other brand POIs.

2 Likes
9

The community debated the specificity of name=* ad nauseam a couple years ago. At the time, a mapper sought unsuccessfully to establish a global norm that chargers, restaurants, and anything else should have a name=* unique to the system it belongs to. Some other mappers feel name=* should be omitted unless it’s distinct from any other tag on the feature – another example of tagging minimalism that creates problems in other respects. Ultimately, OSM ends up with a pragmatic, nuanced compromise between these positions.

My perspective on this is limited, since I’m not an electric vehicle owner and don’t have experience with this kind of infrastructure. What I’ve noticed is that EV charging stations fall into an uncanny valley among points of interest: sort of street furniture, sort of a destination in its own right, depending on how you look at it. That affects expectations about whether it should have a name=* based on brand=*.

Fixtures like fire hydrants, public telephones, information kiosks, and ATMs are generally unnamed with rare exceptions. Data consumers have struggled with whether to label ATMs based on other keys besides name=*. I could see an argument for omitting name=* from an unremarkable charging station in a parking lot, particularly a private one for customers or employees. But many of the ones we’re mapping are intended to be destinations. In rural areas, they’re even signposted at freeway exits just like gas stations and other motorist amenities.

The U.S. has a very strong convention to always include name=* on a gas station. Most chain gas stations will have a name=* that matches the brand=* of either the station itself or the attached store. Any deviation from this practice would need to be well-justified, for example if the gas station is a landmark and its signs promote a unique location name unusually prominently. The point of this naming convention is for each name=* tag to be consistent with end user expectations. To the extent that the names follow a predictable formula, it’s only because of real world expectations and usage, not for the purpose of database consistency.

Either way, I see no reason to follow a name=[City] [Brand] naming formula unless the brand is adamant that this is how you’re supposed to call each charger. The format in a store locator listing isn’t particularly strong evidence of that. Confusingly, “San Jose Supercharger” all but tells the consumer that it’s the Supercharger in San José. Yet there are enough in the city and others like it that this couldn’t possibly be a useful naming scheme either in the real world or in OSM. name=[Brand] seems like the more likely naming scheme, since it potentially distinguishes the charging station from others very close by that are part of other systems.

6 Likes
Charging Stations names vs. Petrol Station names
10

And even then I would only consider it.

11

Sounds reasonable. If there is somewhat of an agreement or official decision, I would like to do the revert of the name= myself. I still made a lot of improvements and fixed a LOT of weird data points along the way and wish that no blind revert removes that work, or even worse, lose data of superchargers I surveyed myself. I fixed so many issues created by those two imports as well as well as old, weirdly and wrongly tagged superchargers.

But I think that would still leave quite some chargers in different state. I didn’t really touch all charger names and many had a name {location} Supercharger before (can’t quantify yet though). Especially since that Tesla Supercharger naming convention was even documented in the wiki when that part still existed:

image
image718×252 19.2 KB

Further, if that’s decided, I would like to get that into iD. Not sure where I would. I tried to get some related changes regarding sockets into that but eventually gave up.

12

Ah, I remember that one fondly… I wonder if gileri ever figured out the name of his local McDonald’s, if it wasn’t just “McDonald’s”.

I’ve parsed through a few of @stillhart’s changes in Canada and didn’t notice any where the name had been changed to name=[Town] Supercharger, but I can imagine that people do actually call many Supercharger locations “the [Town] Supercharger”.

I also don’t own an electric car so my stakes in this discussion are pretty low, but looking at the Tesla Canada website I see that the locations are named “[Town], [Two-letter province abbreviation]” in their list, and that’s how the URLs for each location’s webpage are constructed, e.g. “JasperABsupercharger” for the “Jasper, AB” location in Jasper, Alberta.

stillhart has taken to adding the webpage URL as ref:tesla=*, which seems superfluous to me; it’s just a duplication of the URL. See for example that Supercharger location in Jasper, which is tagged (among many tags):

short_name=Tesla
ref:tesla=JasperABsupercharger
website=https://www.tesla.com/findus/location/supercharger/JasperABsupercharger

I know stillhart didn’t add it, but the “Tesla” short name seems very, very spurious. I doubt very much that anyone calls this “Tesla”. Can anyone imagine a conversation like so:

“Where’re you going to charge your car?”
“Oh, I’m going to charge it at [the] Tesla.”

:face_with_raised_eyebrow: I can’t either…

It seems that the US locations have a number in the URL instead. If this number isn’t present anywhere on site, I would similarly question the ref:tesla tag’s relevance if it’s just duplicating the number at the end of the location webpage’s URL.

All of this said, I could very reasonably imagine people calling Supercharger locations by town name. E.g.:

“Where’re you going to charge your car?”
“I’m going to charge it at the Jasper [Tesla] Supercharger.”

… seems to me like perfectly plausible parlance. Similar to gas stations, e.g. “I’m going to fill up at the Jasper Petro-Canada”. (That said the location in OSM is tagged “Petrocan-Jasper East”… :face_with_spiral_eyes: ) However, this nomenclature quickly falls apart in bigger urban areas with many Supercharger locations. “Jasper Tesla Supercharger” could work because there’s just the one and only location in Jasper; there are 21 locations in San Jose alone? My entire province doesn’t have that many locations. :rofl:

I’m a fan of using branch, which it sounds like many of the American locations already do.

13

Additionally, Tesla tried to change their naming away from [Location] Supercharger to [Business] Supercharger. So [ Mövenpick Hotel ] Supercharger, which expectedly gave a backslash too and they had to revert. Now it shows like this.

image
image1206×788 141 KB

In my interactions I do hear the chargers being called [location] supercharger. But that falls slowly apart in bigger cities and due to the ever more dense Supercharger network. It will stay likely valid in places with no more than 1 Charger.

14

why you think that we have such convention?

15

Because name=*Official Name* Supercharger was written in the Wiki for 6 years including multiple tables full of examples where in the official name is the location in a dedicated article about Tesla Charging Stations as well as lots of OSM nodes matching that. Further, in which about every official name of Tesla is a location and not some kind of unique unrelated name like ‘Map3000 Supercharger’. To add, where iD validates for [anything|Tesla] Supercharger.

Wiki in April 2025:

Wiki in April 2019:

16

There are probably a few exceptions, which are Tesla locations by Tesla themselves, such as the Tesla Diner which is probably locally be called Tesla Diner Supercharger. Los Angeles Supercharger would be pretty wrong here, which I agree on.

image
image1090×632 68 KB

17

On the other hand, the one[1] McDonald’s in Gasoline Alley is “the Gasoline Alley McDonald’s”, and everyone in Alberta knows where it’s at, yeah? People can always informally qualify the name of a POI by containing or nearby place names, just to be helpful, but that doesn’t necessarily make it part of the name=* either.

The wiki could be wrong or outdated. The talk page has complaints about the page’s naming recommendations going back several years. I would put more credence on actual usage.

Personally, I also find “San Jose Supercharger” awkward because the city limits form a very irregular shape. An outlying Supercharger may technically lie just within the San Jose city limits, but this factoid is irrelevant to anyone who needs to interact with the charger, hardly even observable.

Of course, Tesla can do anything they want with their chargers’ names. If they follow Apple or Marriott naming practices, then so would we. But if our only basis for putting the city name in name=* is that this part of the address appears as a subtitle in an online store locator, as a fallback for when they don’t have a more specific name, then I’ve got a Walmart Supercenter to sell you. A fallback that makes sense in a business listing (where the brand is a given but the location needs clarification) isn’t necessarily a fallback that makes sense on a map (which needs to call out the brand).

  1. OK, I know, there are two: northbound and southbound. But still. ↩︎

3 Likes
18

I need a some time, but I will first go for a revert of the duplicates names, hence large cities and what was criticized here. I won’t touch names which haven’t been changed by me.

1 Like
19

It’s obvious this was not an import, and nobody claimed it was an import. What you have done is called a mechanical edit. All imports are mechanical edits, but not all mechanical edits are imports. There were far too many changes done in a short time to have been done without automation assistance and far too many to have time to look at every object. That’s why it’s called a mechanical edit.

It’s easy to program breaks to make it look like a human is doing edits. That’s called gaming the system.

I don’t see much of that in the US, particularly the western US. Maybe you see it elsewhere, but that’s beside the point. The point here is the renaming and that you didn’t follow ANY of the guidelines in the Automated Edits code of conduct. For all we know a 10m-tall Odin appeared to you in a dream and said, “Tomorrow you will be King.” Then you woke up and said, “I don’t have to follow any rules. I am King!”

Combining your last comment with your recent edit history, and the far slower rate of reversion compared to the mechanical edit, it appears that you have no intention of reverting all your name changes, only the ones where there are several Superchargers in a single city. Maybe you think we will go away if you wait long enough?

With the exception of Colorado and Wyoming (which I have already reverted), I therefore call on the Data Working Group to revert all your name changes (and the insertion of the redundant ref:tesla=*) in North America (Canada, US, Mexico).