There is a small but meaningful difference between being given the next song and choosing where to look for it. A recommendation feed usually emphasizes the first. Radiooooo suggests the second by organizing musical exploration around geography and time.
I like the question that interface asks. Where do I want to go, and when? It is easy to answer without knowing an artist, genre or track title. The visitor supplies a direction rather than a search term. That makes the archive approachable before the visitor has acquired its vocabulary.
A map is not a neutral index, of course. A country contains many scenes, languages and overlapping traditions. A decade is a rough container for changes that don't happen on schedule. A small selection of recordings can't represent everything made in either category. The interface works best as a doorway into a collection, not as a claim that the collection defines a place.
That limitation can become a useful design feature if the experience makes its choices visible. Information about the performer, recording and contributor gives the visitor a route beyond the initial selection. A surprising song can become a question about an artist or a scene rather than remaining an anonymous example of a country.
The structure also changes the meaning of randomness. Shuffle across an entire catalog can feel arbitrary. Shuffle within a chosen place and period has a boundary that the listener understands. The system retains the ability to surprise, while the listener retains some authorship over the excursion.
This seems applicable to many collections. A film archive could begin with a location and a production period. A visual archive could begin with a material or a printing process. The useful axis isn't necessarily the one that maximizes predicted engagement. It can be the one that gives a person an interesting question to ask.
I don't think every music interface needs a map. The more general lesson is about pacing. A deliberate choice, followed by enough space to encounter its result, creates a different kind of attention from an endless queue. Sometimes discovery improves when the interface asks for a small decision and then gets out of the way.