Post

I'm having so much fun with this dataset. For some reason I'm doing a lot of analysis of fast food places. Did you know Istanbul has the highest density of fast food restaurants, with Burger King being #1?M

There are genuinely 349 Burger Kings in Istanbul!?

I mean, append "…in the foursquare POI dataset, according to queries I wrote as fast as possible often with unvetted sql from ChatGPT" to everything I wrote, so… probably not.

> Gists automatically render GeoJSON woahh

Really cool!… but aren’t the egress fees on 10GB of free data is going to add up quick? A single query across the full dataset could cost them $0.50 (at a very conservative $0.05/GB)

Why duckdb

Because it can run queries directly against Parquet files over HTTP without needing to download the whole file - it uses HTTP range requests to fetch just the bits of the file it needs to answer the question

I guess DuckDB + S3 combination would slap for deep learning telemetry and plotting

Now that their data is open-source, what value proposition does whatever is left of Foursquare still have?

FourSquare is in the business of data syndication - to Apple, Alexa, etc and to anyone else that needs it. So this dataset is only like a point-in-time export.

So they won't update the open sourced version? Or do they hope for more updates, now that it is open source?

It’s amazing they did this. Suddenly anyone can build an app with geotagging locations. I for one would love to add venues to my Bluesky posts.

There is already a Foursquare API for geotagging, that can be used to do this. Twitter used it in the pre-Elon era. If @bsky.app wants to do this, we would be happy to work with them.

While I got you here, for the love of god, please recompile Swarm app with a modern iOS version and give us a decent photo picker 🙏