There’s a noble history within OSM of people saying “mappers will never want to map X”. “Non-notable trees” was an example, and when people started mapping trees beyond “notable” ones confusion ensued because the people who had mapped “notable” ones couldn’t tell which ones were “notable” any more. This, I suspect, falls into the same category.
I’m not a user of these things but have mapped the things sticking out of the ground that you can connect a car to. Near me they tend to serve 2 car parking spaces (so each one has capacity=2) and, now that there are more of them, they tend to occur together. The “thing sticking out of the ground” is analogous to a petrol pump (which I’ve never bothered mapping in garage forecourts, but some people do) and the “places where they occur together” is analogous to “amenity=fuel” - the refueling area of the garage forecourt.
If some people are using “amenity=charging_station” for the “thing sticking out of the ground” and some for the “places where they occur together”, then people like me, who try and make maps from this data, are going to need some way of telling the two apart. As Simon noted above it isn’t “capacity”, because each “thing sticking out of the ground” might have a capacity of >1 - as I said, near me, it’s often 2. It also surely isn’t the object type within OSM (see my earlier comment).
Anything related to physical appearance is probably out because of the variance between different examples of the same sort of thing - I’m saying “thing sticking out of the ground” and Simon’s saying “wallboxes”, so neither of those is ideal. I really don’t care what we end up using, just that it needs to not be confusing and not already in use for something else.
I’m not following your argument why capacity isn’t sufficient. Does anyone map individual gas pumps? If so, that seems overzealous also. There are apps and websites that are way ahead of OSM for mapping charging stations & I don’t believe any of them map individual stations.
As mentioned further up
Not all chargers are alike
I have been to sites with 3 different chargers that all have the same socket but all deliver different outputs. To be able to map variety of chargers it is important to have them as separate nodes.
Or there are different access restrictions (certain chargers reserved for car sharing services)
Also not having 20 different ref numbers on a node makes it easier to link live data to and OSM object.
Yes people do. You might not, but the tagging scheme needs to support the use case. A good assumption for everything in OSM is that things will always get more detailed over time in ways that maybe weren’t even considered initially. A good tagging scheme doesn’t have to consider every little detail from the get go, but it should take into account that it will be extended and become more detailed in the future.
Hmm did I create more confusion? Instead of boring everybody with more long winded explanations, I would recommend watching
PS: if you look at the Tesla Supercharger here OpenStreetMap those are the actual “chargers” that have two dispensers (the things that are mapped) each. These are version 2 super chargers which is relevant as the charging speed/power will be split between cars that are charging on dispensers that are connected to the same charger.
Maybe I’m missing it, but If you’re referring to post 85 above, showing a Tesla charge location, that does show up as only 1 node on the plugshare website, not individual plugs. I see the same thing on chargehub, openchargemap, electify*. Perhaps it’s semantics, it is one location/map point, with several or many items of data. Are you saying is that doesn’t work within the OSM data structure as a node, too much information for a node?
I’m a new EV owner & I appreciate the effort to get on top of this. I see only 1 charge_station on OSM in my local area, & plugshare has more than 12.
What would be changed about this one: Node: Tesla Supercharger (6565092683) | OpenStreetMap
As far as the name of the individual plugs, how about “plug”
I think they let thing is that it’s important to know how many change points there are at a site, whether they are AC or DC, socket type (less relevant as we standardise on Type 2 in my home country), and charging speed (kW seems to be the metric). You can do this by mapping the individual charge points, or adding the details somehow to the site. As Andy said, some people will want to map individual chargepoints.
“We don’t map individual fuel pumps”
It might help if we did. That way I might have not turned up to a “fuel station” in Germany to discover it was just for aggicultural use with just 1 pump! Driving on to the site with 8 pumps was what I needed.
Let’s see if there could be a compromise. There is some resistance against introducing a new site feature, and the fact that the term station is being used IRL both for a location and for an individual charge point is creating some confusion. How about this variant:
Redefine amenity=charging_station slightly as a “place/location to charge electric vehicles” and which can represent one or a group of charge points. It may be mapped with a node, an area around the charge points or a relation with charge point members (user decides). A relation could be used for dispersed charge points, for example at different corners of a shopping mall or along a street.
For a group (more than one charge point), each physical device (cabinet, pole etc) may be mapped with man_made=charge_point. It is optional to map the individual charge points in a group.
The same, existing tags are available for both features. For amenity=charging_stations: The tag socket:<type>=* will get how many sockets are available at the location, and socket:<type>:output will get the maximum output in kW available for each socket type at the location, in line with current tagging practice.
This will hopefully solve a number of discussed issues:
Search querys would list only the amenity=charging_station features, avoiding 20+ search results for each charge point.
Data consumers would only have to deal with amenity=charging_station (but could also zoom into each charge point at a location, if desired).
The two features, a location and a charge point, corresponds with what is usually displayed in the apps of charging operators (all of the 5 operators which I use).
There is no need to determine exactly what is a site vs an individual charge point other than just to be aware of individually mapped charge points for a group.
The term “charge point” should not be controversial - it has already been used in the industry for several years, for example the Open Charge Point Protocol (OCPP).
Limited need for retagging. Based on the analysis earlier in this thread, there are only 2-7000 cases with groups of individually mapped charge points. 10-40.000 groups of charge points are already mapped with only one node.
Well for starters some of them are sockets In Europe the type 2 AC “chargers” typically don’t have cables with plugs.
But more importantly a dispenser can have multiple connections, here you will find for example Type 2 CCS Combo and Chademo on the same dispenser, or for example version 2 superchargers will have the Tesla plug and a Type 2 CCS Combo. Last Monday I found two AC “chargers” with a type 1 and a type 2 plug (I suspect the installation is oldish and the type 1 plug has never been used).
Now in 5 years or so, I expect that anything else than Type 2 and Type 2 CCS Combo will have vanished completely but till then I’m assuming that multiple connectors will remain a thing. Note that this is also a reason why overall socket/plug counts are not really helpful as they don’t indicate how many simultaneous sessions with a connector of a specific type are supported. .
As has already been pointed out it seems as if a lot (if overpass is to be believed ~24’000 in Europe alone) of those 40’000 are a single or a pair of type 2 AC “chargers” that are not part of a larger “site” so the numbers don’t really support what you are implying.
Unrelated to the argument, but given that numbers are being discussed (24k and 40k in the most recent post) let’s not forget that this is tiny competed to the future. For example:
Shell aims to install 50,000 ubitricity on-street EV charge posts across the UK by 2025
No doubt the other big players will have big plans too, and that’s just one market. Globally the final number is going to be insanely high.
Hence it’s so important that we find a way to agree now at this early stage before it becomes a harder challenge.
Of all the changes proposed, this feels like the easiest tag to add and it gives mappers a clear way to indicate that what is mapped is a single physical asset (charge point, dispenser, wallbox).
How about we move forward with that tag first while still discussing any other changes we may want?
I have provided the detailed breakdown earlier in this thread. Of the amenity=charger_station which are not part of a cluster there are 4658 with capacity=1 , 30454 with capacity=2 and 11205 with capacity=3 and higher. None of these would need any retagging.
Hello everyone,
Thank you for all the feedback. I will be quite busy the next 2 or so weeks and probably won’t have time to work on this proposal. I hope this is not an issue for anyone.
If someone else would like to take over editing the wiki page, please let me know.
So, can man_made=charge_point appear on its own, nowhere near an amenity=charging_station? NKA’s suggestions read to me like an individual charger that isn’t part of a larger site should be mapped as a charging_station with capacity=1/2, not as a charge_point?
And are any changes to capacity:charging proposed? Should people place a charging_station or a charge_point inside the amenity=parking way? What about private ones, do they still qualify as a charging_station?
Well that is the thought behind the question I asked @RobJN
If charge_points can exist stand alone, aren’t they an amenity? And how do they differ from charging_station to start with? Doesn’t this simply create a 2nd tag for the same thing in the end?
In Europe the type 2 AC “chargers” typically don’t have cables with plugs.