Living area estimation using aerial imagery

A computer-implemented method for estimating a living area of a structure, the computer-implemented method comprising:

• determining an accuracy of property data originating from one or more data sources using a trained machine learning model;
• obtaining the property data from a first data source in the one or more data sources having a highest determined accuracy;
• obtaining a request for the living area of the structure;
• processing one or more aerial images corresponding to the structure;
• estimating a first living area of the structure based on the processing of the one or more aerial images;
• determining a property group to which the structure belongs based on the obtained property data;
• modifying the first living area of the structure by a ratio difference corresponding to the property group to which the structure belongs;
• determining a confidence score of the modified first living area;
• determining that the confidence score of the modified first living area is less than a threshold score;
• determining an adjustment metric for the structure based on a portion of the property data corresponding to the structure;
• adjusting the modified first living area using the adjustment metric to form a second living area;
• determining a second confidence score of the second living area;
• and
• transmitting, to a user device, either the modified first living area or the second living area based on the confidence score and the second confidence score.

A non-transitory, computer-readable medium comprising computer-executable instructions for estimating a living area of a structure, wherein the computer-executable instructions, when executed by a computer system, cause the computer system to:

• determine an accuracy of property data originating from one or more data sources using a trained machine learning model;
• obtain the property data from a first data source in the one or more data sources having a highest determined accuracy;
• obtain a request for the living area of the structure;
• process one or more aerial images corresponding to the structure;
• estimate a first living area of the structure based on the processing of the one or more aerial images;
• determine a property group to which the structure belongs based on the obtained property data;
• modify the first living area of the structure by a ratio difference corresponding to the property group to which the structure belongs;
• determine a confidence score of the modified first living area;
• determine that the confidence score of the modified first living area is less than a threshold score;
• determine an adjustment metric for the structure based on a portion of the property data corresponding to the structure;
• adjust the modified first living area using the adjustment metric to form a second living area;
• determine a second confidence score of the second living area;
• and
• transmit, to a user device, either the modified first living area or the second living area based on the confidence score and the second confidence score.