Systems and methods for improving driver safety using uplift modeling

1. A computer implemented method comprising:

receiving, by one or more processors, a first set of telematics data collected by one or more sensors associated with a vehicle driven by a driver;

analyzing, by one or more processors, the first set of telematics data to determine a driving tip is to be provided to the driver;

generating, by one or more processors, a treatment for providing the driving tip to the driver, the treatment being generated using a treatment generation model and including a plurality of treatment facets, the plurality of treatment facets comprising a channel of delivery, a timing of delivery, and a tip content,

wherein the treatment generation model utilizes uplift modeling techniques;

wherein the treatment generation model includes a plurality of options for each treatment facet of the plurality of treatment facets;

wherein the treatment generation model determines a set of drivers having similar characteristics as the driver;

wherein the treatment generation model generates a likelihood of selection for each option of the plurality of options for each treatment facet of the plurality of treatment facets, the likelihood of selection for each option being generated by combining a respective weight associated with the option based upon an effectiveness of the option for drivers within the set of drivers, and a respective confidence score associated with the option based upon a sample size of treatments containing the option provided to the drivers with in the set of drivers; and

wherein the treatment generation model selects an option from the plurality of options for each treatment facet of the plurality of treatment facets based upon the likelihood of selection to generate the treatment;

in accordance with the treatment, providing, by one or more processors, the driving tip to the driver;

after providing the driving tip, receiving, by one or more processors, a second set of telematics data collected by the one or more sensors associated with the vehicle driven by the driver;

determining, by one or more processors, an effectiveness of the driving tip; and

updating, by one or more processors, the treatment generation model by adjusting the respective weight associated with each option of the plurality of options for each treatment facet of the plurality of treatment facets based upon the determined effectiveness using the uplift modeling techniques.

2. The computer implemented method of claim 1 , wherein:

the treatment generation model includes a plurality of possible channels of delivery, timings of delivery, and tip contents; and

the treatment generation model associates each of the possible channels of delivery, timings of delivery, and tip contents with an initial score.

3. The computer implemented method of claim 2 , further comprising:

determining, by one or more processors, the initial score for each of the possible channels of delivery, timings of delivery, and tip contents based upon demographic data of the driver and an effectiveness of each of the possible channels of delivery, timings of delivery, and tip contents for other drivers having similar demographic data.

4. The computer implemented method of claim 3 , wherein the other drivers having similar demographic data are located within a geographic range of the driver.

5. The computer implemented method of claim 2 , wherein:

the treatment generation model associates each of the possible channels of delivery, timings of delivery, and tip contents with a corresponding confidence level based upon a corresponding sample size associated with a number of treatments having the possible channels of delivery, timings of delivery, or tip content provided to other drivers having similar demographic data as the driver.

6. The computer implemented method of claim 5 , wherein generating the treatment comprises:

assigning, by one or more processors, weights to the plurality of possible channels of delivery, timings of delivery, and tip contents based upon respective initial scores and confidence levels; and

selecting, by one or more processors, a particular channel of delivery, a particular timing of delivery, and a particular tip content.

7. The computer implemented method of claim 1 , wherein determining the effectiveness of the driving tip comprises:

comparing, by one or more processors, the first set of telematics data to the second set of telematics data to determine whether the driver complied with the driving tip.

8. The computer implemented method of claim 7 , wherein updating the treatment generation model comprises:

determining, by one or more processors, that the driver complied with the driving tip; and

increasing, by one or more processors, each of the respective weight associated with the selected option for each treatment facet of the plurality of treatment facets utilized in the treatment.

9. The computer implemented method of claim 7 , wherein updating the treatment generation model comprises:

determining, by one or more processors, that the driver did not comply with the driving tip; and

decreasing, by one or more processors, each of the respective weight associated with the selected option for each treatment facet of the plurality of treatment facets utilized in the treatment.

10. The computer implemented method of claim 1 , wherein the treatment includes a tone at which the tip content is provided to the driver.

11. A computer system comprising:

one or more sensors associated with a vehicle;

one or more processors;

one or more transceivers adapted to communicate with the vehicle driven by a driver;

a non-transitory program memory coupled to the one or more processors and storing executable instructions that, when executed by the one or more processors, cause the computer system to:

receive, via the one or more sensors, a first set of telematics data associated with the vehicle;

analyze the first set of telematics data to determine a driving tip is to be provided to the driver;

generate, using a treatment generation model, a treatment for providing the driving tip to the driver, the treatment including a plurality of treatment facets, the plurality of treatment facets comprising a channel of delivery, a timing of delivery, and a tip content,

wherein the treatment generation model utilizes uplift modeling techniques;

wherein the treatment generation model includes a plurality of options for each treatment facet of the plurality of treatment facets;

wherein the treatment generation model determines a set of drivers having similar characteristics as the driver;

wherein the treatment generation model generates a likelihood of selection for each option of the plurality of options for each treatment facet of the plurality of treatment facets, the likelihood of selection for each option being generated by combining a respective weight associated with the option based upon an effectiveness of the option for drivers within the set of drivers, and a respective confidence score associated with the option based upon a sample size of treatments containing the option provided to the drivers within the set of drivers; and

wherein the treatment generation model selects an option from the plurality of options for each treatment facet of the plurality of treatment facets based upon the likelihood of selection to generate the treatment;

in accordance with the treatment, provide the driving tip to the driver;

after providing the driving tip, receive, via the one or more sensors, a second set of telematics data associated with the vehicle driven by the driver;

determine an effectiveness of the driving tip; and

update the treatment generation model by adjusting the respective weight associated with each option of the plurality of options for each treatment facet of the plurality of treatment facets based upon the determined effectiveness using the uplift modeling techniques.

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

determine the respective weight associated with each option of the plurality of options for each treatment facet of the plurality of treatment facets based upon demographic data of the driver and an effectiveness of each option of the plurality of options for each treatment facet of the plurality of treatment facets for other drivers having similar demographic data.

13. The computer system of claim 12 , wherein the other drivers having similar demographic data are located within a geographic range of the driver.

14. The computer system of claim 11 , wherein:

the treatment generation model associates each option of the plurality of options for each treatment facet of the plurality of treatment facets with a confidence level based upon a corresponding sample size associated with a number of treatments having the option and provided to other drivers having similar demographic data as the driver.

15. The computer system of claim 11 , wherein to determine the effectiveness of the driving tip, the instructions, when executed by the one or more processors, cause the system to:

compare the first set of telematics data to the second set of telematics data to determine whether the driver complied with the driving tip.

16. The computer system of claim 15 , wherein to update the treatment generation model, the instructions, when executed by the one or more processors, cause the system to:

determine that the driver complied with the driving tip; and

increase each of the respective weight associated with the selected option for each treatment facet of the plurality of treatment facets utilized in the treatment.

17. The computer system of claim 15 , wherein to update the treatment generation model, the instructions, when executed by the one or more processors, cause the system to:

determine that the driver did not comply with the driving tip; and

decrease each of the respective weight associated with the selected option for each treatment facet of the plurality of treatment facets utilized in the treatment.

18. A non-transitory computer readable storage medium storing processor-executable instructions, that, when executed, cause one or more processors to:

receive a first set of telematics data associated with a vehicle, the first set of telematics data being collected by one or more sensors associated with the vehicle;

analyze the first set of telematics data to determine that a driving tip is to be provided to the driver;

generate, using a treatment generation model, a treatment for providing the driving tip to the driver, the treatment including a plurality of treatment facets, the plurality of treatment facets comprising a channel of delivery, a timing of delivery, and a tip content,

wherein the treatment generation model utilizes uplift modeling techniques;

wherein the treatment generation model includes a plurality of options for each treatment facet of the plurality of treatment facets;

wherein the treatment generation model determines a set of drivers having similar characteristics as the driver;

wherein the treatment generation model generates a likelihood of selection for each option of the plurality of options for each treatment facet of the plurality of treatment facets, the likelihood of selection for each option being generated by combining a respective weight associated with the option based upon an effectiveness of the option for drivers within the set of drivers, and a respective confidence score associated with the option based upon a sample size of treatments containing the option provided to the drivers within the set of drivers; and

wherein the treatment generation model selects an option from the plurality of options for each treatment facet of the plurality of treatment facets based upon the likelihood of selection to generate the treatment;

in accordance with the treatment, provide the driving tip to the driver;

after providing the driving tip, receive a second set of telematics data collected by the one or more sensors associated with the vehicle driven by the driver;

determine an effectiveness of the driving tip; and

update the treatment generation model by adjusting respective weight associated with each option of the plurality of options for each treatment facet of the plurality of treatment facets based upon the determined effectiveness using the uplift modeling techniques.

19. The computer implemented method of claim 1 , wherein the tip content includes a money amount, a likelihood of damage or injury, a comparison to other drivers, or an indication of performance over time.

20. The computer system of claim 11 , wherein the tip content includes a money amount, a likelihood of damage or injury, a comparison to other drivers, or an indication of performance over time.