Predicting hiring priorities

1. A method, comprising:

with training data relating to previous hiring activity of companies, training both a first machine learning model to produce a first set of scores representing future hiring volumes and a second machine learning model to produce a second set of scores representing future hiring growth, the training further comprising:

generating clusters of hiring features associated with identified patterns of hiring activity using k-means clustering,

generating a set of values of metrics related to hiring activity over a first time period,

generating a set of labels from the set of values of metrics over a second time period, and

generating a normalized ranking of the hiring features based on the set of labels;

selecting hiring features characterizing hiring activity by a company for a set of titles;

applying, by one or more computer systems, the first trained machine learning model to the hiring features to produce a first set of scores representing future hiring volumes by the company for the set of titles;

applying the first trained machine learning models to a set of company features to produce a second set of scores representing future hiring growth;

generating, by the one or more computer systems, a first ranking of the set of titles by the first set of scores and the second set of scores;

applying the second trained machine learning model to the hiring features to produce a third set of scores representing future hiring growth by the company for the set of titles;

applying the second trained machine learning model to a set of company features to produce a fourth set of scores;

generating, by the one or more computer systems, a second ranking of the set of titles by the third set of scores and the fourth set of scores;

outputting at least a portion of the first ranking as a first prediction of the future hiring volumes by the company for the set of titles; and

outputting at least a portion of the second ranking as a second prediction of the future hiring growth by the company for the set of titles.

2. The method of claim 1 , wherein the set of company features comprise at least one of:

a company identifier;

a company size;

a company name;

a company age;

a company type;

a company industry; and

a company location.

3. The method of claim 1 , further comprising:

aggregating metrics for characterizing the hiring activity of a set of companies for the set of titles into a number of clusters;

selecting, from the metrics, the hiring features and labels for the first and second machine learning models based on patterns associated with the number of clusters; and

inputting the hiring features and the labels for the set of companies as training data for the first and second machine learning models.

4. The method of claim 3 , further comprising:

generating the values of the hiring features from values of the metrics collected over a first time period; and

generating the labels from values of the metrics collected over a second time period following the first time period.

5. The method of claim 4 , wherein generating the values of the hiring features from the values of the metrics comprises:

converting a numerical feature for the company and a title into a proportion of the numerical feature within the company.

6. The method of claim 4 , wherein generating the values of the hiring features from the values of the metrics comprises:

converting a range of a numerical feature to a normalized range of 0 to 1.

7. The method of claim 3 , wherein aggregating the metrics for characterizing the hiring activity of the set of companies for the set of titles into the number of clusters comprises:

generating the number of clusters from the metrics for the set of companies associated with a company size and an industry.

8. The method of claim 1 , further comprising:

generating a recommendation related to hiring by the company based on the first or third sets of scores.

9. The method of claim 8 , wherein the recommendation comprises at least one of:

a hiring strategy;

a hiring budget; and

a characteristic of a talent pool.

10. The method of claim 1 , wherein the hiring features comprise at least one of:

a number of candidates that received hiring messages from the company;

a number of hiring messages sent by the company;

a number of hiring messages accepted by the candidates;

a number of hiring messages rejected by the candidates;

a number of hires made by the company;

a number of applications for jobs at the company;

a number of applicants for jobs at the company;

a number of jobs posted by the company; and

a growth in a hiring feature from a first period to a second period.

11. A system, comprising:

one or more processors; and

memory storing instructions that, when executed by the one or more processors, cause the system to:

with training data relating to previous hiring activity of companies, train both a first machine learning model to produce a first set of scores representing future hiring volumes and a second machine learning model to produce a second set of scores representing future hiring growth, the training further comprising:

generate clusters of hiring features associated with identified patterns of hiring activity using k-means clustering,

generate a set of values of metrics related to hiring activity over a first time period,

generate a set of labels from the set of values of metrics over a second time period, and

generate a normalized ranking of the hiring features based on the set of labels;

select hiring features characterizing hiring activity by an entity for a set of talent pools;

apply a first trained machine learning model to the hiring features to produce a first set of scores representing future hiring growth by the entity for the set of talent pools;

apply the first trained machine learning model to a set of company features for the company to produce a second set of scores representing future hiring growth;

generate a first ranking of the set of talent pools by the first set of scores and the second set of scores;

output at least a portion of the first ranking as a first prediction of the future hiring growth by the entity;

apply a second trained machine learning model to the hiring features to produce a third set of scores representing future hiring volumes by the entity for the set of talent pools;

apply the second trained machine learning model to the set of company features for the company to a fourth set of scores;

generate a second ranking of the set of talent pools by the third set of scores and the fourth set of scores; and

output at least a portion of the second ranking as a second prediction of the future hiring volumes by the entity.

12. The system of claim 11 , wherein the memory further stores instructions that, when executed by the one or more processors, cause the system to:

aggregate metrics for characterizing the hiring activity of a set of entities for the set of talent pools into a number of clusters;

select, from the metrics, the hiring features and labels for the first and second machine learning models based on patterns associated with the number of clusters; and

input the hiring features and the labels for the set of entities as training data for the first and second machine learning models.

13. The system of claim 12 , wherein the memory further stores instructions that, when executed by the one or more processors, cause the system to:

generate the values of the hiring features from values of the metrics collected over a first time period; and

generate the values of the labels from values of the metrics collected over a second time period following the first time period.

14. The system of claim 13 , wherein generating the values of the hiring features from the values of the metrics comprises at least one of:

converting a numerical feature for the entity and a talent pool into a proportion of the numerical feature within the entity; and

converting a range of a numerical feature to a normalized range of 0 to 1.

15. The system of claim 11 , wherein the set of talent pools comprises at least one of:

a title;

a location;

an industry;

a seniority;

a function;

a company; and

a company size.

16. The system of claim 11 , wherein the entity comprises at least one of:

a company;

a location;

an industry; and

a company size.

17. A non-transitory computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method, the method comprising:

with training data related to previous hiring activity of companies, training both a first machine learning model to produce a first set of scores representing future hiring volumes and a second machine learning model to produce a second set of scores representing future hiring growth, the training further comprising:

generating clusters of hiring features associated with identified patterns of hiring activity using k-means clustering,

generating a set of values of metrics related to hiring activity over a first time period,

generating a set of labels from the set of values of metrics over a second time period, and

generating a normalized ranking of the hiring features based on the set of labels;

selecting hiring features characterizing hiring activity by a company for a set of titles;

applying a first trained machine learning model to the hiring features to produce a first set of scores representing future hiring volumes by the company for the set of titles;

applying the first trained machine learning model to the company features to produce a second set of scores representing future hiring volumes;

applying a second trained machine learning model to the hiring features to produce a third set of scores representing future hiring growth by the company for the set of titles;

applying the second trained machine learning model to the set of company features to produce a fourth set of scores representing future hiring growth;

generating a first ranking of the set of titles by the first and second set of scores and a second ranking of the set of titles by the third and fourth set of scores;

outputting at least a portion of the first ranking as a first prediction of the future hiring volumes by the company; and

outputting at least a portion of the second ranking as a second prediction of the future hiring growth by the company.