A non-transitory computer-readable medium comprising instructions that, when executed by a hardware processor of a device, cause the device to perform operations comprising:
- accessing content data of a first media asset and a set of input cues associated with the first media asset;
- for each individual timestamp of a series of timestamps of the first media asset:
- identifying, by an element parser that is implemented using the hardware processor, a plurality of different-mode elements of the first media asset based on the content data and the set of input cues, the element parser comprising a first plurality of machine learning models that each identify a different mode of element, the plurality of different-mode elements corresponding to the individual timestamp of the first media asset;
- assigning, by an element analyzer that is implemented using the hardware processor, a set of weight values to the plurality of different-mode elements for the individual timestamp based on a priority rule, the element analyzer comprising a second plurality of machine learning models that each generate a weight value for a different mode of element;
- and
- identifying one or more relevant different-mode elements from the plurality of different-mode elements based on a ranking of the set of weight values;
- analyzing a plurality of relevant elements for the series of timestamps using a machine learning process to generate a set of classified features based on a set of tags for the series of timestamps, each tag corresponding to an event or scene in the content data of the first media asset, the plurality of relevant elements comprising the one or more relevant elements for each of the individual timestamps;
- using a machine learning algorithm to generate a content genome of the first media asset based on the set of classified features, the content genome comprising instruction code that uniquely identifies and organizes one or more flaggable events in view of cultural attributes associated with a plurality of different geographical regions;
- causing a graphical user interface to be displayed on a client device, the graphical user interface comprising:
- an audience rating timeline that comprises a plurality of representative user interface elements for multiple different audience groups in the plurality of different geographical regions, the plurality of representative user interface elements comprising:
- a first individual representative user interface element that indicates whether the content data at a given timestamp is suitable for consumption by a first audience age group in a first geographical region of the plurality of different geographical regions based on the content genome;
- and
- a second individual representative user interface element that indicates whether the content data at the given timestamp is suitable for consumption by a second audience age group in a second geographical region of the plurality of different geographical regions based on the content genome;
- and
- a graphical user element that provides rating exception counts for the plurality of geographical regions;
- and
- performing an automated method for capturing one or more time-based markers of the first media asset that represent a duration of one or more objectionable scenes in the content data, at least one of the one or more objectionable scenes comprising a scene relating to violence, profanity, or drugs, at least one time-based marker corresponding to the given timestamp, wherein the one or more time-based markers are provided to a streaming platform to enable a skip scene feature with respect to the first media asset;
- accessing a cultural attribute from a cultural attribute graph, the cultural attribute being associated with an individual geographical region;
- using a machine learning algorithm to generate a culture-specific content gene of the first media asset based on the set of classified features and the cultural attribute;
- identifying a first title of the first media asset based on the plurality of different-mode elements of content data;
- matching, using a machine learning algorithm, the culture-specific content gene with a category in a pre-determined content genome graph;
- based on the matching, generating a first culture-specific content genome of the individual geographical region for the first title of the first media asset, the first culture-specific content genome comprising a first cluster of culture-specific content genes;
- identifying a second title of a second media asset based on a plurality of different-mode elements of content data of the second media asset;
- generating a second culture-specific content genome of the individual geographical region for the second title of the second media asset, the second culture-specific content genome comprising a second cluster of culture-specific content genes;
- determining a degree of overlap between the first cluster of culture-specific content genes and the second cluster of culture-specific content genes, both the first cluster of culture-specific content genes and the second cluster of culture-specific content genes being associated with the individual geographical region;
- and
- based on the determining of the degree of overlap, associating the first media asset with the second media asset for media asset recommendations in the individual geographical region.