Hybrid parser and machine-learning-based extraction of commercial flight bookings and matching to empty leg shared charters

A non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, cause an online shared charter platform to perform operations comprising:

• accessing a data object from an external data platform, the data object being associated with a user of the online shared charter platform;
• identifying, from the accessed data object, a commercial flight booking of the user;
• attempting, by a set of template-based parsers executed by the online shared charter platform, to extract a set of parser-based structured commercial flight attributes from the commercial flight booking;
• responsive to determining that the set of template-based parsers failed to produce a validated set of parser-based structured commercial flight attributes, inputting the identified commercial flight booking into a machine-learned model trained to output a set of model-based structured commercial flight attributes, wherein the machine-learned model is guided by a structured data template defining a machine-readable schema for causing the set of model-based structured commercial flight attributes to conform to the structured data template;
• storing the parser-based or model-based structured commercial flight attributes as a structured commercial flight dataset in a datastore accessible to a matching engine of the online shared charter platform;
• matching, by the matching engine, the structured commercial flight dataset with an available empty leg flight based on geographic and temporal matching constraints;
• and
• transmitting a notification comprising the available empty leg flight to a client device associated with the user in response to determining that the available empty leg flight satisfies a matching criterion relative to the structured commercial flight dataset.

A system, comprising:

• at least one memory;
• and
• at least one processor coupled with the at least one memory, the at least one memory storing code comprising instructions that, when executed by the at least one processor, cause an online shared charter platform to perform operations comprising:
• accessing a data object from an external data platform, the data object being associated with a user of the online shared charter platform;
• identifying, from the accessed data object, a commercial flight booking of the user;
• attempting, by a set of template-based parsers executed by the online shared charter platform, to extract a set of parser-based structured commercial flight attributes from the commercial flight booking;
• responsive to determining that the set of template-based parsers failed to produce a validated set of parser-based structured commercial flight attributes, inputting the identified commercial flight booking into a machine-learned model trained to output a set of model-based structured commercial flight attributes, wherein the machine-learned model is guided by a structured data template defining a machine-readable schema for causing the set of model-based structured commercial flight attributes to conform to the structured data template;
• storing the parser-based or model-based structured commercial flight attributes as a structured commercial flight dataset in a datastore accessible to a matching engine of the online shared charter platform;
• matching, by the matching engine, the structured commercial flight dataset with an available empty leg flight based on geographic and temporal matching constraints;
• and
• transmitting a notification comprising the available empty leg flight to a client device associated with the user in response to determining that the available empty leg flight satisfies a matching criterion relative to the structured commercial flight dataset.