Show HN: I discovered roads in the US across > 1000 themes | Hacker News Reader

Show HN: I discovered roads in the US across > 1000 themes

road-about-a-theme.pinedesk.biz

13 points91awebsi11 comments

This interactive dataviz I created lets you explore about ~1k different themes I computed by cross referencing popular Wikidata articles against OSM roads.

Open articleView on HN

The framing is a bit deceptive, saying that these roads are "named after" these things, despite the clarification in the about that they may have no relation. This whole thing is a bit busted, the about on each "chapter" only sort of describes what a category entails. Obscure shares the same about as famous.

This feels like a project that was done without much care for the details, considering how many themes are pure nonsense. No acknowledgement of AI in how this was done but it sure smells of it.

Agreed. I chose Robot Names, and it gave me a Mars Orbital Mission for Mom Lane, and Cygnus was a Deke Slayton reference. It seemed more about space and satellites than robots with personalities. And the thumbnail number (over 9000!) did not match the clicked result. And the segmented sections were hard to follow and navigate.

For another example: one "theme" listed is "Walking Dead episodes". The streets which come up are ones with names like "Faith" or "Family", which happen to also be names of episodes of that show. There is clearly no actual connection here.

ouch, and also thank you for calling out the busted.

An LLM titles each theme, but otherwise this is all derived by checking wikidata entities that are 'most read' (according to qrank), then choosing from a set of sparql queries I put together (members of group X, places within Y, etc), for what produces the most results, and then string/substring/bm25 matching across all the streets in the country.

I think the trash smell is too generously allowing an association both in my sparql queries (what to search for), and whats considered a match :/

I wonder too if explaining how a street matched (following the wikidata graph) would make the nonsensical less offending.

Busted?

You really think people wouldn't notice the slop? Did you even look at the results before hitting "submit"?

Streets named after famous people, "Don John", it thinks it's named after Donald (John) Trump.

"David Robert Court", oh "David Robert" is the first 2 names of David Beckham.

Please... I'm going to flag this submission because of your lying and stupidity. "I discovered"? Go fuckkkk offfff. (Sorry for breaking HN rules, but this is fucking obnoxious)

Interesting idea. 1,000 is too many to browse for fun. Maybe expend some human effort and pick a few more interesting ones to feature.

What I saw on the homepage made little sense: "We searched every named street in America for Communicators. 108 names turned up. A rare one we found is Big Bang Court in Washoe County, NV."

Please do all of HN a favor and flag this slop to death... otherwise, every other moron with quarter-assed "AI"-assisted idea is going to think "Oh this is HN-worthy".

There are not 6875 roads named after Lana Del Rey. "Del Rey" means "of the king" in Spanish and is just a popular fancy-sounding name to use as a street name in subdivisions.

The Cheshire Cat example is more on target and was amusing to learn about.

My favorite juxtaposition of street names exists in Brooklyn, Indiana, a small town of ~2600 people.

On the south side you face a stark choice: when you leave Home Avenue, you can either head west to Church Street, or east to Hooker Street.

An intentional set that I'm amused by: https://maps.app.goo.gl/GQ5xKJwKtBqut7Cx9

This Street, That Street, and The Other Street in Halifax.

This is a really fun use of cross-referencing datasets. Curious how you decided which Wikidata themes were worth including.

Reply on news.ycombinator.com