Systems and methods of establishing correlative relationships between geospatial data features in feature vectors representing property locations
View Patent ↗In an illustrative embodiment, an automated system and method develops customized feature vectors from geospatial information system (GIS) information. The system and method may obtain GIS features located within a predetermined boundary surrounding a property location and store the GIS features within a feature vector. The system can add amplifying data features accessed from external data sources that expand a knowledge scope of GIS feature(s). Amounts of correlation between the GIS features and associated amplifying data features within the feature vector and a property value of the property location can be identified using at least one data model trained with a data set including types of GIS features of the customized feature vector of the property location.
1 . A method for automatically appraising property value in view of features located within a predetermined vicinity of a property, the method comprising:
determining, by at least one processor, a predetermined boundary surrounding a property location;
obtaining, by the at least one processor from a geospatial information system (GIS) information data set, a plurality of sets of feature data comprising a respective set of feature data for each respective GIS feature of a plurality of GIS features located within or overlapping with the predetermined boundary surrounding the property location, wherein
the plurality of GIS features comprises one or more of a set of amenity features, a set of transportation features, a set of land use features, a set of economic features, or a set of public utility features,
each GIS feature of the plurality of GIS features comprises a respective location identified using respective geographic coordinates representing a line, a point, or a polygon, and
each set of feature data comprises a respective type of the respective GIS feature, and a respective identifier of the respective GIS feature;
storing, to a non-transitory computer readable storage region, the plurality of sets of feature data in a feature vector of the property location;
accessing, by the at least one processor, a plurality of amplifying data features collected from one or more external data sources, each amplifying data feature related to at least one respective GIS feature of the plurality of GIS features, wherein
each amplifying data feature of the plurality of amplifying data features expands a knowledge scope of the at least one respective GIS feature, and
each amplifying data feature of the plurality of amplifying data features is accessed using at least one of the identifier of the at least one respective GIS feature or the geographic coordinates of the at least one respective GIS feature;
adding, by the at least one processor, each of the plurality of amplifying data features to the corresponding set of feature data in the feature vector of the property location; and
for each respective set of feature data of the plurality of sets of feature data of the feature vector, identifying, by the at least one processor using at least one machine learning data model, an amount of correlation between the respective set of feature data and a property value of the property location, wherein
each machine learning data model of the at least one machine learning data model was trained with a respective data set corresponding to at least a portion of a plurality of types of GIS features represented in the plurality of GIS features located within or overlapping with the predetermined boundary surrounding the property location, wherein
the respective data set corresponding to one or more types of GIS features of the plurality of types of GIS features comprises a set of amplifying data obtained from at least one respective external data source of the one or more external data sources.
2 . The method of claim 1 , wherein the property location is a parcel of land.
3 . The method of claim 1 , wherein determining the predetermined boundary comprises applying a bounding box to the property location.
4 . The method of claim 3 , wherein obtaining the plurality of sets of feature data comprises:
determining that a number of an initial plurality of GIS features obtained is fewer than a minimum threshold number of GIS features; and
increasing a size of the bounding box.
5 . The method of claim 1 , wherein the plurality of amplifying data features comprises, for each set of feature data of a portion of the plurality of sets of feature data, respective ratings data obtained from one or more ratings data sources of the one or more external data sources.
6 . The method of claim 5 , wherein:
a first GIS feature type of the one or more types of GIS features of a given set of feature data of the plurality of sets of feature data represents a school; and
a corresponding amplifying data feature of the plurality of amplifying data features comprises quality rating data obtained from a school ratings data source of the one or more ratings data sources.
7 . The method of claim 1 , wherein the plurality of amplifying data features comprises, for each set of feature data of a portion of the plurality of sets of feature data, respective catastrophic weather information obtained from at least one of a weather zone data source, a catastrophic modeling data source, or a land topology data source of the one or more external data sources.
8 . The method of claim 7 , wherein each respective amplifying data feature of at least one amplifying data feature of the plurality of amplifying data features comprises a likelihood of natural disaster or a shoreline position prediction.
9 . The method of claim 1 , further comprising:
for each GIS feature of the plurality of GIS features, calculating, by the at least one processor, a respective distance from the respective GIS feature to the property location;
wherein each set of feature data of the plurality of sets of feature data further comprises the respective distance.
10 . The method of claim 9 , wherein calculating the respective distance comprises, for each GIS feature of the plurality of GIS features identified using respective geographic coordinates representing the line or the polygon, identifying a closest point in the line or the polygon to one of the property location or a bounding box surrounding the property location.
11 . The method of claim 1 , further comprising:
receiving, by the at least one processor, a request submitted by a remote computing device for a property value estimate of the property location;
wherein the amount of correlation is identified responsive to receiving the request.
12 . The method of claim 11 , further comprising, based on the request, adding, by the at least one processor, one or more additional sets of features to the feature vector, each additional set of features being customized to one or more preferences associated with the request.
13 . The method of claim 1 , further comprising updating, by the at least one processor, the feature vector of the property location responsive to receiving another GIS information data set comprising one or more GIS features within the predetermined boundary of the property location.
14 . A system for automatically appraising property value in view of features located within a predetermined vicinity of a property, the system comprising:
a non-transitory computer readable storage region configured to store a plurality of feature vectors;
a plurality of machine learning data models, each machine learning data model of the plurality of machine learning data models trained with a respective data set corresponding to at least a portion of a plurality of types of geospatial information system (GIS) features, wherein
the respective data set corresponding to one or more types of GIS features of the plurality of types of GIS features comprises a set of amplifying data obtained from at least one respective external data source of one or more external data sources; and
at least one processor configured to determine a predetermined boundary surrounding a property location,
obtain, from a GIS information data set, a plurality of sets of feature data comprising a respective set of feature data for each respective GIS feature of a plurality of GIS features located within or overlapping with the predetermined boundary surrounding the property location, wherein
the plurality of GIS features comprises one or more of a set of amenity features, a set of transportation features, a set of land use features, a set of economic features, or a set of public utility features,
each GIS feature of the plurality of GIS features comprises a respective location identified using respective geographic coordinates representing a line, a point, or a polygon, and
each set of feature data comprises a respective type of the respective GIS feature, and a respective identifier of the respective GIS feature,
store, to the non-transitory computer readable storage region, the plurality of sets of feature data in a respective feature vector of the plurality of feature vectors, wherein the respective feature vector is associated with the property location,
access a plurality of amplifying data features collected from at least one of the one or more external data sources, each amplifying data feature related to at least one respective GIS feature of the plurality of GIS features, wherein
each amplifying data feature of the plurality of amplifying data features expands a knowledge scope of the at least one respective GIS feature, and
each amplifying data feature of the plurality of amplifying data features is accessed using at least one of the identifier of the at least one respective GIS feature or the geographic coordinates of the at least one respective GIS feature,
add each of the plurality of amplifying data features to the corresponding set of feature data in the feature vector of the property location, and
for each respective set of feature data of the plurality of sets of feature data of the feature vector, identify, using at least one machine learning data model of the plurality of machine learning data models, an amount of correlation between the respective set of feature data and a property value of the property location, wherein
the at least one machine learning data model comprises, for each respective GIS feature type of a subset of the plurality of types of GIS features, one or more machine learning data models trained with the respective GIS feature type, wherein
the subset of the plurality of types of GIS features is represented in the feature vector.
15 . The system of claim 14 , wherein:
determining the predetermined boundary comprises applying a bounding box to the property location; and
obtaining the plurality of sets of feature data comprises:
determining that a number of an initial plurality of GIS features obtained is fewer than a minimum threshold number of GIS features, and
increasing a size of the bounding box.
16 . The system of claim 14 , wherein the plurality of amplifying data features comprises, for each set of feature data of a portion of the plurality of sets of feature data, respective ratings data obtained from one or more ratings data sources of the one or more external data sources.
17 . The system of claim 14 , wherein the plurality of amplifying data features comprises, for each set of feature data of a portion of the plurality of sets of feature data, respective catastrophic weather information obtained from at least one of a weather zone data source, a catastrophic modeling data source, or a land topology data source of the one or more external data sources.
18 . The system of claim 14 , wherein the at least one processor is further configured to:
for each GIS feature of the plurality of GIS features, calculate a respective distance from the respective GIS feature to the property location;
wherein each set of feature data of the plurality of sets of feature data further comprises the respective distance.
19 . The system of claim 14 , wherein the at least one processor is further configured to:
receive a request submitted by a remote computing device for a property value estimate of the property location;
wherein the amount of correlation is identified responsive to receiving the request.
20 . The system of claim 14 , wherein the at least one processor is further configured to update the feature vector of the property location responsive to receiving another GIS information data set comprising one or more GIS features within the predetermined boundary of the property location.