Station library creation for a media service
A machine may form all or part of a network-based system configured to provide media service to one or more user devices. The machine may be configured to define a station library within a larger collection of media files. In particular, the machine may access metadata that describes a seed that forms the basis on which the station library is to be defined. The machine may determine a genre composition for the station library based on the metadata. The machine may generate a list of media files from the metadata based on a relevance of each media file to the station library. The machine may determine the relevance of each media file based on a similarity of the media file to the genre composition of the station library as well as a comparison of metadata describing the media file to the accessed metadata that describes the seed.
1. A method comprising:
accessing a station descriptor profile generated from a seed defining a station set;
selecting a plurality of candidate media files;
computing a similarity score for each candidate media file, the similarity score including a measure of similarity between the corresponding candidate media file and the station descriptor profile; and
machine-generating the station set to include a subset of the plurality of the candidate media files based at least on the similarity score for each candidate media file,
wherein the station descriptor profile includes one or more focus genre profiles and each candidate media file includes a file genre profile, the one or more focus genre profiles and the file genre profile each specifying respective multiple genres and a weight assigned to each genre, the weight assigned to each genre indicating a percentage weighting of the genre relative to the other genres in the respective multiple genres specified for the corresponding focus genre profile or file genre profile, and
computing the similarity score for each corresponding candidate media file includes:
computing one or more focus-level similarity scores by comparing, for each focus genre profile in the station descriptor profile: (a) the respective multiple genres and the corresponding weights of the focus genre profile, and (b) the respective multiple genres and the corresponding weights of the file genre profile of the corresponding candidate media file; and
selecting the highest focus-level similarity score to be the similarity score for the corresponding candidate media file.
2. The method of claim 1 , wherein the plurality of candidate media files is selected according to a relationship of each candidate media file to the seed.
3. The method of claim 1 , wherein the seed defining the station set is based on a seed artist and each focus genre profile in the station descriptor profile relates to the seed artist.
4. The method of claim 3 , wherein each candidate media file is associated with a candidate artist, and the method further comprises determining artist relation data indicating how the candidate artist for each candidate media file is related to the seed artist.
5. The method of claim 4 , wherein the artist relation data includes, for each candidate media file, a relationship type and a relationship weight, the relationship weight measuring a strength of the relationship type between the seed artist and the candidate artist for the corresponding candidate media file.
6. The method of claim 4 , wherein the plurality of candidate media files are selected according to the artist relation data.
7. The method of claim 4 , wherein the station set is machine-generated to include the subset of the plurality of the candidate media files based additionally on the artist relation data.
8. The method of claim 1 , further comprising allocating at least one portion of the station set to a respective focus genre profile in the station descriptor profile.
9. The method of claim 8 , further comprising assigning at least one of the candidate media files to the at least one allocated portion of the station set based on the highest focus-level similarity score for the at least one candidate media file.
10. The method of claim 1 , wherein the station descriptor profile specifies one or more additional attributes for the station set, the method further comprises determining further comparisons between the station descriptor profile and the plurality of candidate media files based on the one or more additional attributes, and the station set is machine-generated to include the subset of the plurality of the candidate media files based additionally on the further comparisons.
11. The method of claim 10 , wherein the one or more additional attributes includes mood characteristic.
12. The method of claim 1 , wherein the method further comprises determining a popularity metric for each candidate media file, and the station set is machine-generated to include the subset of the plurality of the candidate media files based additionally on the popularity metric of each candidate media file.
13. The method of claim 1 , wherein the method further comprises determining a language attribute for each candidate media file, and the station set is machine-generated to include the subset of the plurality of the candidate media files based additionally on the language attribute of each candidate media file.
14. The method of claim 1 , wherein the method further comprises determining temporal data for each candidate media file, and the station set is machine-generated to include the subset of the plurality of the candidate media files based additionally on the temporal data of each candidate media file.
15. The method of claim 1 , wherein the method further comprises determining user activity data for each candidate media file, and the station set is machine-generated to include the subset of the plurality of the candidate media files based additionally on the user activity data of each candidate media file.
16. The method of claim 1 , wherein the method further comprises receiving one or more constraints from at least one of a user device or an editor device, and removing one or more of the candidate media files from the station set to satisfy the one or more constraints.
17. The method of claim 1 , wherein each focus genre profile in the station descriptor profile is associated with a respective artist.
18. A non-transitory machine-readable storage medium comprising instructions that, when executed by one or more processors of a machine, cause the machine to perform operations comprising:
accessing a station descriptor profile generated from a seed defining a station set;
selecting a plurality of candidate media files;
computing a similarity score for each candidate media file, the similarity score including a measure of similarity between the corresponding candidate media file and the station descriptor profile; and
machine-generating the station set to include a subset of the plurality of the candidate media files based at least on the similarity score for each candidate media file,
wherein the station descriptor profile includes one or more focus genre profiles and each candidate media file includes a file genre profile, the one or more focus genre profiles and the file genre profile each specifying respective multiple genres and a weight assigned to each genre, the weight assigned to each genre indicating a percentage weighting of the genre relative to the other genres in the respective multiple genres specified for the corresponding focus genre profile or file genre profile, and
computing the similarity score for each corresponding candidate media file includes:
computing one or more focus-level similarity scores by comparing, for each focus genre profile in the station descriptor profile: (a) the respective multiple genres and the corresponding weights of the focus genre profile, and (b) the respective multiple genres and the corresponding weights of the file genre profile of the corresponding candidate media file; and
selecting the highest focus-level similarity score to be the similarity score for the corresponding candidate media file.
19. A system comprising:
one or more processors of a machine; and
a machine-readable medium storing instructions that, when executed by the one or more processors of a machine, cause the machine to perform operations comprising:
accessing a station descriptor profile generated from a seed defining a station set;
selecting a plurality of candidate media files;
computing a similarity score for each candidate media file, the similarity score including a measure of similarity between the corresponding candidate media file and the station descriptor profile; and
machine-generating the station set to include a subset of the plurality of the candidate media files based at least on the similarity score for each candidate media file,
wherein the station descriptor profile includes one or more focus genre profiles and each candidate media file includes a file genre profile, the one or more focus genre profiles and the file genre profile each specifying respective multiple genres and a weight assigned to each genre, the weight assigned to each genre indicating a percentage weighting of the genre relative to the other genres in the respective multiple genres specified for the corresponding focus genre profile or file genre profile, and
computing the similarity score for each corresponding candidate media file includes:
computing one or more focus-level similarity scores by comparing, for each focus genre profile in the station descriptor profile: (a) the respective multiple genres and the corresponding weights of the focus genre profile, and (b) the respective multiple genres and the corresponding weights of the file genre profile of the corresponding candidate media file; and
selecting the highest focus-level similarity score to be the similarity score for the corresponding candidate media file.