A non-transitory computer-readable medium storing instructions for adjusting a dynamic discount, the instructions upon execution by one or more processors of a computing system, cause the computing system to perform one or more processes including:
- receiving an enrollment request from a user associated with a vehicle for a vehicle insurance policy;
- identifying a set of user characterizations;
- determining a group of similar existing users based at least in part upon the set of user characterizations;
- determining an initial discount based at least in part upon a set of discount rates associated with the group of similar existing users;
- issuing the initial discount to the user for the vehicle insurance policy at a first time;
- collecting, at least in part by one or more sensors, trip data associated with a set of vehicular trips operated by the user with the vehicle, wherein the trip data, as collected, includes telematics data and travel condition data;
- determining one or more gap periods in the trip data, as collected, during which the telematics data was not collected;
- predicting expected trip data for a trip, including gap-period travel condition data of the travel condition data, for the one or more gap periods, at least in part by:
- weighing the travel condition data differently based upon similarities between one or more respective travel conditions of (a) the gap-period travel condition data and (b) non-gap-period trip data of the travel condition data;
- assigning a date and a day of a week to the telematics data, as collected outside of the one or more gap periods;
- weighing the telematics data differently based on (a) a respective date assigned to the telematics data and (b) a respective day of the week assigned to the telematics data, comprising:
- (1) weighing the telematics data assigned to a particular date of a particular week based on an amount of telematics data collected during the particular date in comparison to respective amounts of telematics data collected during other days of the particular week to produce a first score for the telematics data assigned to the particular date, and wherein the telematics data assigned to the particular date of the particular week has a higher weight than the telematics data assigned to a different date of the particular week when more telematics data is collected during the particular date of the particular week than on the different date of the particular week;
- (2) weighing the telematics data assigned to the particular date of the particular week (wherein the particular date occurs on a particular day of the week) based on how recent the particular date is compared to the particular day of the week of other weeks that are assigned to the telematics data to produce a second score for the telematics data assigned to the particular date, wherein the telematics data assigned to the particular date occurring on the particular day of the week has a higher weight than the telematics data assigned to a different date occurring on a same particular day of the week of an earlier week;
- (3) weighing the telematics data assigned to the particular date of the particular week based on the amount of telematics data collected during the particular date of the particular week in comparison to the respective amounts of telematics data collected during other dates of any week to produce a third score for the telematics data assigned to the particular date, wherein the telematics data assigned to the particular date has a higher weight than the telematics data assigned to the other dates when more telematics data is collected during the particular date than during the other dates;
- and
- repeating (1), (2), and (3) for the telematics data on other particular dates of the particular week and of the other weeks;
- determining an average driving behavior each day of the week by averaging (for the particular day of the week) the first score, the second score, and the third score for the telematics data assigned to all days of the particular day of the week;
- and
- predicting the expected trip data based on the travel condition data, as weighed, and the average driving behavior, as determined, for the particular day of the week that the trip is to occur;
- dynamically updating the initial discount based at least in part upon the trip data, as collected, and the expected trip data as the trip data are collected;
- after dynamically updating the initial discount, determining whether the initial discount has dropped below a minimum discount threshold;
- and
- generating a notification to the user for encouraging improvement in driving behaviors of the user after determining that the initial discount has dropped below the minimum discount threshold.