HN Explorer

Show HN: Placekey – An ID for Every Place

24 points by arjie ↗ HN
Hi, HN, I'm Roshan and I'm part of the team that's launching the Placekey today at https://placekey.io. It is a free and open universal standard identifier we built so that data pertaining to places can be shared across organizations easily.

A Placekey is a short identifier that has Address and POI sections detailing What is at a place and a section describing Where on the globe the place is. Eventually, there will be a Placekey for every place in the world, even those without postal addresses. US geospatial data from a large array of partners https://www.placekey.io/partners can already be associated and cross-referenced using Placekey. By allowing every place to be represented by a canonical string, data scientists and researchers can easily join datasets and share geospatial information with each other.

Generating a Placekey will always be free, whether you do so using https://placekey.io or the API available at https://dev.placekey.io. And storing a Placekey is permitted for any reason, whether you want to cache it, use it to exchange information with someone else, or to fulfill any other purpose you can think of.

Finally, any requests you make to the Placekey API are completely private. They will be used to improve Placekey's matching and resolution, but will never be shared with or sold to third parties.

Some more details are in our:

* FAQ - https://placekey.io/faq

* API Documentation - https://docs.placekey.io/

Try generating one at https://placekey.io/ No signup required. Let us know your thoughts!

13 comments

[ 3.0 ms ] story [ 43.8 ms ] thread
evanbarry
Glad to see this on here. Was planning to attend their launch event tomorrow.
mtmail
The FAQ reads "What3Words is not an open standard". Placekey might be open, how the values are encoded is published in the Technical specs, but the database in central and owned by one company.

From the PDF "In order to maintain the What part of a Placekey, SafeGraph maintains databases of addresses and POI. Incoming addresses and POI are either matched against pre-existing places in our databases, or they are assigned new Placekeys in our database."

"Should a business ping the Placekey API with a massive number of requests—in the millions per month, for example—we may invite you to help us cover those computing costs." I understand the intent, it's providing an API and that has a cost. But isn't the risk the limits are set lower and lower in the future since there is no competition to generating placekeys?

arjie
Thanks for reading, mtmail. By design, once you have a key you can just pass it around. It isn't encoded to you or anything. So as you've appended a placekey to more of your datasets you become more independent from Placekey since you really don't need the API anymore.

As for the cost, our objective is to just run this for free for small users and at-cost for anyone who has substantial usage. We're still in the early stages here so we haven't provided explicit numbers or commitments here but I'm curious to hear what you think. What would reassure you as to our intent?

And for what it's worth, one of our aims is stability so if you want to cache responses for long periods of time, that would be A-OK.

todsacerdoti
This reminds me of when we launched the VAST (https://www.iab.com/guidelines/vast/) standard in online video advertising. Standards lift all boats. Congrats on the launch as this is a big step forward for location based products.
arjie
Thanks, Tod. The fact that data companies also gain from network effects when their datasets can be combined was quite the insight for me when I first encountered it¹. The naive way is to silo off your data, but network effects get you way farther.

¹ https://twitter.com/auren/status/1141010679706476544

rysqui
I work a lot with address and place data. When I tested the API I was surprised how robust it is to name and address variation. It knows "MLK blvd #101" versus "Martin Luther King Blvd STE 101" are the same address. It also is robust to typos or variation in name like "Village Pizza" vs "Village Inn Pizza & Pasta". I also tested some of my favorite edge scenarios like two starbucks across the street with similar addresses, and the API knows that these are different places and does not merge them together. This natural ambiguity is what makes address and place data so challenging, and my tests show that the API is achieving something non-trivial. I'm impressed. The list of launch partners also includes a lot of major platforms (e.g., Esri, Carto). Definitely worth a look.
arjie
Thanks, rysqui! Part of what makes it so effective is the fact that there is a large POI database and search engine behind it and not just a set of addresses or geometries.
trippl7777
I am also attending the launch event tomorrow! Great to see it here! I have worked with the PlaceKeys (n) a bit already and it is amazing to have a standard that is decipherable. Rather than some long string of characters that seemingly have no rhyme or reason, it is nice to have a code that stands for something on the user end! The native inclusion of batch processing made the work flow much much faster as well! Looking forward to the event on October 7th!
daltonlp
Hey, that's great!

Can you help me decipher the difference between the PlaceKeys "zzw-222" and "222-222"?

I'd love to know what those stand for on the user end.

YaBa
Since Google already has Plus Codes, what's new here?