Every edge in the Stell-R graph represents a documented influence connection between ~25,000 artists, pulled from album reviews and music-history writing where one artist's name is explicitly cited in critiques of another's work. Not streaming co-listens, user similarity, or genre tags, just documented influence. My core methodology is at: Badillo-Goicoechea, E. (2025). Modeling Artist Influence for Music Selection and Recommendation: A Purely Network-Based Approach. Harvard Data Science Review
No reviews yetBe the first to leave a review for Stell-R
Maker
📌
Built this just for fun and curiosity as a data scientist + music lover as Stell-R, a music discovery tool based on pure, documented influence lineage ibased on artist reviews, instead of AI / algorithmic user similarity.
Report
How do you handle cases where an artist is cited in critiques of another artist, but the citation is negative or dismissive? Does that still get logged as a direct influence edge, or is there some kind of weighting based on tone?
Report
Maker
@melisatekkol Ah, really good point. Fortunately, as an avid album review reader myself, I can attest that this is quite uncommon (artists being cited explicitly as NOT having influenced the one in question). Rather, almost all reviews are mentioned because both artists share at least some context from the writer's perspective. In any case, I do measure the strength of the association by # times mentioned (so only 1 mention ever counts less than 2+) and also by sonic metrics!
Report
How do you keep the influence edges clean when a review mentions an artist in passing rather than crediting them as an influence? Seems like that distinction matters a lot at this scale.
Report
Maker
@zehra108697 absolutely. currently working on strengthening the edge formation process in order to remove randomness. Any ideas/insights are super welcome!
Report
Looked up a few artists I know and the influence chains actually feel accurate, especially the ones connecting post-rock to krautrock. The cited review excerpts behind each link are a really nice touch for trusting the data.
Report
Maker
@berkrat9 oh, so nice to know this and thanks for trying it out! Stay tuned (I'm constantly improving the pipeline!).
Report
Finally tried tracing a few obscure threads through the graph and was impressed by how cleanly it surfaces the critic-cited lineage rather than vague genre overlap.
How do you handle cases where an artist is cited in critiques of another artist, but the citation is negative or dismissive? Does that still get logged as a direct influence edge, or is there some kind of weighting based on tone?
@melisatekkol Ah, really good point. Fortunately, as an avid album review reader myself, I can attest that this is quite uncommon (artists being cited explicitly as NOT having influenced the one in question). Rather, almost all reviews are mentioned because both artists share at least some context from the writer's perspective. In any case, I do measure the strength of the association by # times mentioned (so only 1 mention ever counts less than 2+) and also by sonic metrics!
How do you keep the influence edges clean when a review mentions an artist in passing rather than crediting them as an influence? Seems like that distinction matters a lot at this scale.
@zehra108697 absolutely. currently working on strengthening the edge formation process in order to remove randomness. Any ideas/insights are super welcome!
Looked up a few artists I know and the influence chains actually feel accurate, especially the ones connecting post-rock to krautrock. The cited review excerpts behind each link are a really nice touch for trusting the data.
@berkrat9 oh, so nice to know this and thanks for trying it out! Stay tuned (I'm constantly improving the pipeline!).
Finally tried tracing a few obscure threads through the graph and was impressed by how cleanly it surfaces the critic-cited lineage rather than vague genre overlap.
@rmeysairmalx6a That is music to my ears (no pun intended!)