Hacker News Reader

Comment by Mordisquitos | Hacker News Reader

Checking on numberplates and bin logos was hardly groundbreaking, as those are exactly where I would have started too. However, I had never heard of Overpass Turbo [0]. Learning of its existence was the only valuable take from this video for me, but well worth it.

[0] https://overpass-turbo.eu/

View on HN

How dare someone mention some common sense tips as a tiny part of a video. Shameful!

Basically, that is the gist of the vid, the rest is more or less common sense, I guess.

A short & catchy variation of "Overpass turbo - a super powerful mining tool for OSM" would be a more appropriate title [0][1]. It exists since 2013 but got a massive boost from the Pokémon Go craze in 2016 [2].

[0]https://wiki.openstreetmap.org/wiki/Overpass_turbo

[1]https://wiki.openstreetmap.org/wiki/Overpass_API

[2]https://www.itechpost.com/articles/34612/20160930/pokemon-go...

And in this case I think it would have been a lot of work to identify the number plate & bin logo if he hadn’t already had enough context to make a probable guess. The source image he was working from looked pretty low-res, that license plate was just a couple white-ish pixels. Like you, I hadn’t heard of overpass turbo and that looks super cool, but the rest of it was “how to confirm what you already mostly know”.

Also I would not have assumed that ChatGPT could generate code for such a niche tool, so that was useful as well.

I mean, it's probably niche if you consider all software ever, but the tool has been around since 2013 at least (and before that under a different name, OSM Server Side Scripting/OSM3S), with a lot of ecosystem tools and documentation (in multiple languages) existing all over the place, I wouldn't say it's very niche in the mapping community.

I agree, GPT3.5 has been working insanely well for me when making overpass turbo queries. I guess there's lots of training data in the dataset for it. Made a quick 10 liner that takes in a bbox & search prompt -> GPT3.5 -> overpass turbo -> geojson

i.e. "banks in my area" a la google maps here: https://i.imgur.com/pSRn1xo.png

In my experience it seems like prompting for older technologies works much better on large LLMs, i guess it makes sense considering that there is probably more crawlable documentation out there

EDIT: someone in this thread linked a tool which seems to do this: https://github.com/rowheat02/osm-gpt

It's a surprising helpful tool in Pokemon Go metagaming. For example, you can use it to determine gyms eligible for EX Raid status (not that that's a thing anymore), find potential micro-nests, or even locate Lake Trio spawn locations.

I'm thinking of a new game: write the shortest Overpass Turbo query that yields, say, 1 result for eg. your country.

One of these I had seen had almost no hints, but it did have a church that was fairly close to a kfc and a dollar general or something. It also kind of “looked” like the southeast, so the dude did a general query on places where those three types of structures inhabited the same general area. It came back with about 7 or 8 results in the entire southeast which could then be manually checked so it seemed like being able to query the map did the big lifting there.

This comment has no valuable take.

I think it does: It saves watching the video by providing a link to a tool that he mentions that many might not have heard of.

Reply on news.ycombinator.com