Music genre map

Introduction

People have always sought to systematize music, breaking it into genres and subgenres. The main method of such systematization was and remains the “historical” method, which can be described as follows:
There is a performer K , who plays music similar to the genre L , but obviously this genre does not fit. The performer himself or the music. the critic, the label, comes up with a new genre for this kind of music, the genre L - LL . And after R years, the LL genre becomes canonical. Or does not, depending on the results of the "test of time." Thus, all musical genres can be combined into a graph, in which the genre L is closer to the root, and LL - to the leaves.

This method is historical, since subgenres always appeared later than the main genres. Those. This method illustrates the development of music, its evolution. But such a graph does not quite correctly illustrate the connections of genres with each other, from a flat musical point of view. Those. from two completely different musical directions, from different eras, subgenres may appear so close that a significant number of musicians will indicate both genres as their “leiba”. But in the “historical” graph they are unlikely to be near.
Moreover, due to the high degree of subjectivity and complexity of the historical method, there is no “single tree of genres”. As far as I am aware, there are only some successful attempts to build such a graph at the level of separate groups of close genres.

There is an alternative method of systematization, which is devoid of these shortcomings. It can be described as follows: There are M performers playing music that falls under 1-4 canonical musical genres. Genres will be connected with an edge in the event that both genres are joined together within the framework of one performer. The more often artists have intersections of both genres, the closer their nodes will be to each other in the final graph (shorter edges). The more often the genre is found in general, the larger its knot will be in relation to the rest.
')
You can see what happened as a result here .

Some observations: Kraut Rock, turned out to be far enough away from the “fatal” constellation, and got into the experimental area of ​​the graph, and Post Grunge turned out to be Metal, not Grunge. Also, not a single genre was in the center, the overall figure of the graph is more like a donut (ring).
I assume that there will be a lot of criticism in the comments. To some, this graph will seem untenable. Take it easy, this graph is just a concept, illustrating one of the variants of the methodology for determining formal relationships between music genres.

Technical part

JavaScript engine “D3” is chosen as an engine. Before this, Protovis and Arbor were tested. But the first one did not cope with the presented amount of data, the second one does not have the best algorithm for branch positioning.

about the project

The source of information about artists and their genres was the Musical Informer . The principle of the service: the user selects artists who are personally interested in him, and when someone of them will have an album or single - the user will receive a notification via E-mail. No quizzes, advertisements, promotions, only notifications of new releases.