Method and system for predicting and indexing real estate demand and pricing

1. A computer-implemented method for predictive modeling, the method comprising:

aggregating, by one or more processors, sample data regarding a plurality of factors associated with employment and geographic location;

scrubbing, by one or more processors, the sample data;

performing, by one or more processors, iterative analysis on the sample data using machine learning to construct a predictive model, wherein the iterative analysis comprises:

randomly dividing, by one or more processors, the sample data into a training subset and a test subset;

applying, by one or more processors, the training subset in machine learning to construct the predictive model;

applying, by one or more processors, the test subset to the predictive model to generate test results;

determining, by one or more processors, whether the test results meet a desired accuracy, and;

responsive to determining that the test results do not meet the desired accuracy, repeating, by one or more processors, the iterative analysis;

populating, by one or more processors using the predictive model, a database with predicted values of a real estate demand for a selected set of predefined geographic regions;

converting, by one or more processors, the predicted values of a real estate demand in the database into percentages of observed values of real estate demand for geographic regions within the selected set over a specified time period to create indices of real estate demand; and

rank ordering, by one or more processors, the geographic regions within the selected set according to their indices of real estate demand.

2. The method according to claim 1 , further comprising:

comparing, by one or more processors, the rank ordering of real estate demand for the selected set of predefined geographic regions to observed real estate demand for the predefined geographic regions over a second specified time period;

aggregating, by one or more processors, updated sample data over the second specified time period; and

updating, by one or more processors, the predictive model using machine learning incorporating the updated sample data for the second specified time period.

3. The method according to claim 1 , wherein categories of the sample data applied to the machine learning to construct the predictive model include:

at least one of:

salary;

total payroll deductions; or

tax filing status; and

at least one of:

job type;

work location;

region of residence;

home values;

rental costs;

available rental properties;

houses for sale;

total houses;

total rental properties;

total retail locations;

available retail locations;

retail rental costs;

employer growth trends;

industry/sector growth trends;

regional employment; or

industry/sector diversity.

4. The method according to claim 1 , wherein the machine learning uses supervised learning to construct the predictive model.

5. The method according to claim 1 , wherein the machine learning uses unsupervised learning to construct the predictive model.

6. The method according to claim 1 , wherein the machine learning uses reinforcement learning to construct the predictive model.

7. A machine learning predictive modeling system, comprising:

a computer system; and

one or more processors running on the computer system, wherein the one or more processors are operable to,

aggregate sample data regarding a plurality of factors associated with employment and geographic location;

scrub the sample data;

perform iterative analysis on the sample data using machine learning to construct a predictive model, wherein the iterative analysis comprises:

randomly dividing the sample data into a training subset and a test subset;

applying the training subset in machine learning to construct the predictive model:

applying the test subset to the predictive model to generate test results;

determining whether the test results meet a desired accuracy, and

responsive to determining that the test results do not meet the desired accuracy, repeating the iterative analysis;

populate, using the predictive model, a database with predicted values of real estate demand for a selected set of predefined geographic regions;

convert the predicted values of real estate demand in the database into percentages of observed values of real estate demand for geographic regions within the selected set over a specified time period to create indices of real estate demand; and

rank order the geographic regions within the selected set according to their indices of real estate demand.

8. The machine learning predictive modeling system according to claim 7 , wherein the one or more processors running on the computer system compare the rank ordering of real estate demand for the selected set of predefined geographic regions to observed real estate demand for said geographic regions over a second specified time period; aggregate updated sample data over the second specified time period; and update the predictive model using machine learning incorporating the updated sample data for the second specified time period.

9. The machine learning predictive modeling system according to claim 7 , wherein the one or more processors comprise aggregated graphical processor units (GPU).

10. The machine learning predictive modeling system according to claim 7 , wherein the machine learning uses supervised learning to construct the predictive model.

11. The machine learning predictive modeling system according to claim 7 , wherein the machine learning uses unsupervised learning to construct the predictive model.

12. The machine learning predictive modeling system according to claim 7 , wherein the machine learning uses reinforcement learning to construct the predictive model.

13. A computer program product for machine learning predictive modeling, the computer program product comprising:

a persistent computer-readable storage media;

program code, stored on the computer-readable storage media, for aggregating sample data regarding a plurality of factors associated with employment and geographic location;

program code, stored on the computer-readable storage media, for scrubbing, by one or more processors, the sample data;

program code, stored on the computer-readable storage media, for performing iterative analysis on the sample data using machine learning to construct a predictive model, wherein the iterative analysis comprises:

randomly dividing the sample data into a training subset and a test subset;

applying the training subset in machine learning to construct the predictive model;

applying the test subset to the predictive model to generate test results;

determining whether the test results meet a desired accuracy, and

responsive to determining that the test results do not meet the desired accuracy, repeating the iterative analysis;

program code, stored on the computer-readable storage media, for populating, using the predictive model, a database with predicted values of real estate demand for a selected set of predefined geographic regions;

program code, stored on the computer-readable storage media, for converting the predicted values of real estate demand in the database into percentages of observed values of real estate demand for geographic regions within the selected set over a specified time period to create indices of real estate demand; and

program code, stored on the computer-readable storage media, for rank ordering the geographic regions within the selected set according to their indices of real estate demand.

14. The computer program product according to claim 13 , further comprising:

program code, stored on the computer-readable storage media, for comparing the rank ordering of real estate demand for the selected set of predefined geographic regions to observed real estate demand for said geographic regions over a second specified time period;

program code, stored on the computer-readable storage media, for aggregating updated sample data over the second specified time period; and

program code, stored on the computer-readable storage media, for updating the predictive model using machine learning incorporating the updated sample data for the second specified time period.

15. The computer program product according to claim 13 , wherein categories of the sample data applied to the machine learning to construct the predictive model include:

at least one of:

salary;

total payroll deductions; or

tax filing status; and

at least one of:

job type;

work location;

region of residence;

home values;

rental costs;

available rental properties;

houses for sale;

total houses;

total rental properties;

total retail locations;

available retail locations;

retail rental costs;

employer growth trends;

industry/sector growth trends;

regional employment; or

industry/sector diversity.

16. The computer program product according to claim 13 , wherein the machine learning uses supervised learning to construct the predictive model.

17. The computer program product according to claim 13 , wherein the machine learning uses unsupervised learning to construct the predictive model.

18. The computer program product according to claim 13 , wherein the machine learning uses reinforcement learning to construct the predictive model.