IP-API.com discussion

[gepus]

https://www.iplocation.net/

Error margin from my exact location is aprox 18 miles. Desktop (IPv4) with geo disabled in the browser.

[Moonchild]

https://www.iplocation.net/

Error margin from my exact location is aprox 18 miles. Desktop (IPv4) with geo disabled in the browser.

Your point being....?

[New Tobin Paradigm]

Some people do care because they are too lazy to enter a zip code to localize results like in the local store chain example or w/e. I always run with geo.enabled set to false personally but that is me.

ALSO, don't forget, laptops are mobile devices with wifi that run desktop software as well as the couple of x86-based tablets out there

[gepus]

https://www.iplocation.net/

Error margin from my exact location is aprox 18 miles. Desktop (IPv4) with geo disabled in the browser.

Your point being....?

Error margins respectively accuracy of the tracking.

[Moonchild]

Error margins respectively accuracy of the tracking.

I still don't understand what point you're trying to make.
Desktop geolocation accuracy will always depend on several factors:

• The size and geographic spread of the IP pool your IP address is in.
• The accuracy of the geoip lookup tables

That being said, the given accuracy is normally enough for web services to select appropriate local services for you.

Without GPS and drawing purely on IP address, exact accuracy is not possible.

Google works around this by keeping a database of not only IP addresses but also local/wifi networks; if Google's service is used it will only be more accurate if private network information is included in the request (which e.g. Firefox and Chrome do by default). This is the reason why people have started saying that geolocation is an invasion of privacy but that is ONLY if you use a service that takes more than just your public IP as input. We/ip-api.com don't.

[Samson]

Safety issues aside, these services can't even make up their collective minds as to which part of Utah I'm in at my desktop PC. So the accuracy is total garbage.

[Moonchild]

Safety issues aside, these services can't even make up their collective minds as to which part of Utah I'm in at my desktop PC. So the accuracy is total garbage.

I repeat:

Desktop geolocation accuracy will always depend on several factors:

• The size and geographic spread of the IP pool your IP address is in.
• The accuracy of the geoip lookup tables

[Samson]

Yes, I got that, but it's being touted as some kind of solution to a problem when it's wildly inaccurate in a lot of cases and seems to be of little use on a desktop PC when it can't even pinpoint what part of a state you're in.

So while I don't see it as unsafe and am not going to concern myself with trying to block it, it also doesn't seem like it's worth anyone's time to implement it either.

[Moonchild]

I'll rephrase it for you: the accuracy of IP-based geolocation depends very heavily on your ISP and how they distribute their address space among clients.

[Samson]

I understood you the first time. You don't seem to be getting that some of us think the data is worthless because of that very problem and thus not worth the time to implement it. Which makes me and others wonder why any company thinks it's useful at all if this is a common thing.

[Moonchild]

Just because your accuracy isn't better than "what state am I in" doesn't make it worthless and unnecessary to have implemented...
Think a little broader than your own backyard for a minute. (pun intended)

Also, the accuracy provided is often all a website ever needs to cater to something local, e.g. selecting your state's branch office for a world-wide or national chain. IP-API.com in my opinion strikes a good balance between providing service and not providing too much information (which would be a privacy concern).
If you find it worthless, you can disable geolocation in the browser. Servers will then likely do the lookup themselves based on your IP or simply refuse to provide local services. Your choice.