US-12625901-B2 - Machine learning and language model-assisted geospatial data analysis and visualization

US12625901B2US 12625901 B2US12625901 B2US 12625901B2US-12625901-B2

Abstract

Methods and systems for site prospecting includes the operations of: receiving a site request indicating a required use for a site; generating a plurality of capacity scores corresponding to a plurality of land parcels using a first machine learning model; filtering the plurality of land parcels into a subset of land parcels based on the plurality of capacity scores; and for at least one land parcel in the subset of land parcels: generating a parcel potential description using a first language model based at least in part on geographic information associated with the at least one land parcel; generating a parcel potential score using a second machine learning model based at least in part on the parcel potential description; and presenting the parcel potential description and the parcel potential score.

Inventors

Nikhil Nainani
Gabriel Seite
Sameer Zaheer
Elena Haddad
Dudon Wai

Assignees

Palantir Technologies Inc.

Dates

Publication Date: 20260512
Application Date: 20240516

Claims (20)

1 . A method for site prospecting using machine learning models, the method comprising: receiving a site request indicating a required use for a site; generating a plurality of capacity scores corresponding to a plurality of land parcels by using a first machine learning model; filtering the plurality of land parcels by using the plurality of capacity scores to reduce the plurality of land parcels into a subset of land parcels; and for at least one land parcel in the subset of land parcels, generating a parcel potential description by using a first language model based at least in part on geographic information associated with the at least one land parcel; generating a parcel potential score by applying a second machine learning model on the parcel potential description; generating a parcel feasibility description by using a second language model; and generating a parcel feasibility score by applying a third machine learning model on the parcel feasibility description; wherein the method is performed using one or more processors.
2 . The method of claim 1 , further comprising: presenting the parcel feasibility description and the parcel feasibility score.
3 . The method of claim 1 , further comprising: retrieving feasibility information associated with the site request; wherein the generating a parcel feasibility description comprises generating the parcel feasibility description by applying the second language model to a parcel location of a respective parcel and the feasibility information comprising unstructured text data.
4 . The method of claim 3 , further comprising: receiving an indication of a feasibility data source associated with the site request; wherein the retrieving feasibility information comprises retrieving the feasibility information from the feasibility data source.
5 . The method of claim 3 , wherein the feasibility information is received via a large language model plugin.
6 . The method of claim 1 , wherein at least one of the first language model or the second language model is a large language model.
7 . The method of claim 1 , further comprising: retrieving geographic information associated with the site request, the geographic information comprising unstructured text data; wherein the generating a parcel potential description comprises generating the parcel potential description by applying the first language model to a parcel location of a respective parcel and the one or more geographic requirements and a respective capacity score for the respective parcel.
8 . The method of claim 7 , further comprising: receiving an indication of a geographic data source associated with the site request; wherein the retrieving geographic information comprises retrieving the geographic information from the geographic data source.
9 . The method of claim 7 , wherein the geographic information is received via a large language model plugin.
10 . The method of claim 1 , wherein one or more site parameters include at least one selected from a group consisting of a size, a geographic parameter, a use parameter, a wind parameter, a solar parameter, and a site use.
11 . The method of claim 1 , wherein each land parcel of the plurality of land parcels has a substantially similar size.
12 . The method of claim 11 , wherein the substantially similar size is determined based at least in part on the site request.
13 . The method of claim 1 , wherein the first machine learning model is different from the second machine learning model.
14 . The method of claim 1 , wherein the presenting the parcel potential description and the parcel potential score comprises presenting the parcel potential description and the parcel potential score with a geographic map.
15 . A system for site prospecting using machine learning models, the system comprising: one or more memories having instructions stored therein; and one or more processors configured to execute the instructions and perform operations comprising: receiving a site request indicating a required use for a site; generating a plurality of capacity scores corresponding to a plurality of land parcels by using a first machine learning model; filtering the plurality of land parcels by using the plurality of capacity scores to reduce the plurality of land parcels into a subset of land parcels; and for at least one land parcel in the subset of land parcels, generating a parcel potential description by using a first language model based at least in part on geographic information associated with the at least one land parcel; generating a parcel potential score by applying a second machine learning model on the parcel potential description; generating a parcel feasibility description by using a second language model; and generating a parcel feasibility score by applying a third machine learning model on the parcel feasibility description.
16 . The system of claim 15 , wherein the operations further comprise: presenting the parcel feasibility description and the parcel feasibility score.
17 . The system of claim 15 , wherein the operations further comprise: retrieving feasibility information associated with the site request; wherein the generating a parcel feasibility description comprises generating the parcel feasibility description by applying the second language model to a parcel location of a respective parcel and the feasibility information comprising unstructured text data.
18 . The system of claim 17 , wherein the operations further comprise: receiving an indication of a feasibility data source associated with the site request; wherein the retrieving feasibility information comprises retrieving the feasibility information from the feasibility data source.
19 . The system of claim 15 , wherein at least one of the first language model or the second language model is a large language model.
20 . A method for site prospecting using machine learning models, the method comprising: receiving a site request indicating a required use for a site, and a plurality of geographic areas; generating a plurality of geographic area capacity scores corresponding to the plurality of geographic areas by using a first machine learning model; selecting a geographic area from the plurality of geographic areas based on the plurality of geographic area capacity scores; generating a plurality of land parcel capacity scores corresponding to the plurality of land parcels by using a second machine learning model; filtering the plurality of land parcels by using the plurality of land parcel capacity scores to reduce the plurality of land parcels into a subset of land parcels; for at least one land parcel in the subset of land parcels, generating a parcel potential description by using a first language model based at least in part on geographic information associated with the at least one land parcel; generating a parcel potential score by applying a third machine learning model on the parcel potential description; generating a parcel feasibility description by using a second language model; and generating a parcel feasibility score by applying a fourth machine learning model on the parcel feasibility description; wherein the method is performed using one or more processors.

Description

CROSS REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Application No. 63/503,827, filed May 23, 2023, and U.S. Provisional Application No. 63/532,259, filed Aug. 11, 2023, the disclosures of which are incorporated by reference herein in their entireties for all purposes. TECHNICAL FIELD Certain embodiments of the present disclosure are directed to systems and methods for selecting sites. More particularly, some embodiments of the present disclosure provide systems and methods for selecting sites using machine learning models and/or language models based on data. BACKGROUND A site prospecting process usually begins with a very large geographical area, such as a state or a province, and involves a lot of manual analysis (e.g., of geospatial data) and rule-based workflows to manually narrow down the large geographical area to a handful of sites for prospective development. Such manual processing can be very time-consuming (e.g., months or weeks) and expensive, typically requiring many site visits to select a site for the particular use, and can incur opportunity costs as well as fail to identify good sites due to pre-defined inflexible rules. Computerized systems may be used to record and visualize geospatial data, for example, including wind data such as wind speed, direction and sheer, solar data such as solar irradiance data. Hence it is desirable to improve the techniques for geospatial data analysis and visualization. SUMMARY Certain embodiments of the present disclosure are directed to systems and methods for selecting sites. More particularly, some embodiments of the present disclosure provide systems and methods for selecting sites using machine learning models and/or language models based on data. Disclosed are methods and systems for site prospecting. According to some embodiments, the method includes: receiving a site request indicating a required use for a site; generating a plurality of capacity scores corresponding to a plurality of land parcels using a first machine learning model; filtering the plurality of land parcels into a subset of land parcels based on the plurality of capacity scores; and for at least one land parcel in the subset of land parcels: generating a parcel potential description using a first language model based at least in part on geographic information associated with the at least one land parcel; generating a parcel potential score using a second machine learning model based at least in part on the parcel potential description; and presenting the parcel potential description and the parcel potential score. The method is performed using one or more processors. According to some embodiments, the system includes one or more memories having instructions stored therein and one or more processors configured to execute the instructions and perform operations. The operations include: receiving a site request indicating a required use for a site; generating a plurality of capacity scores corresponding to a plurality of land parcels using a first machine learning model; filtering the plurality of land parcels into a subset of land parcels based on the plurality of capacity scores; and for at least one land parcel in the subset of land parcels: generating a parcel potential description using a first language model based at least in part on geographic information associated with the at least one land parcel; generating a parcel potential score using a second machine learning model based at least in part on the parcel potential description; and presenting the parcel potential description and the parcel potential score. According to some embodiments, the method includes: receiving a site request indicating a required use for a site, and a plurality of geographic areas; generating a plurality of geographic area capacity scores corresponding to a plurality of geographic areas using a first machine learning model; selecting a geographic area from the plurality of geographic areas based on the plurality of geographic area capacity scores; breaking the selected geographic area into a plurality of land parcels; generating a plurality of land parcel capacity scores corresponding to the plurality of land parcels using a second machine learning model; filtering the plurality of land parcels into a subset of land parcels based on the plurality of land parcel capacity scores; for at least one land parcel in the subset of land parcels: generating a parcel potential description using a first language model based at least in part on geographic information associated with the at least one land parcel, generating a parcel potential score using a third machine learning model based at least in part on the parcel potential description; generating a parcel feasibility description using a second language model, and generating a parcel feasibility score using a fourth machine learning model; and presenting at least one selected from a group consisting of: the parcel p