Privacy-safe frequency distribution of geo-features for mobile devices
A mobile device's location and an identifier corresponding to the mobile device is received. The mobile device's location is privatized by mapping it to landmarks proximate to the mobile device's location, storing the proximate landmarks in association with the mobile device's identifier, and discarding the received location data. The proximate landmarks are featurized to generate a model which is used to determine a value of the advertising opportunity corresponding to a target identifier.
1. A method, comprising:
privatizing, by a server, geographic location information, comprising:
for each of a plurality of mobile devices, for each of a plurality of locations:
receiving a location and an identifier corresponding to a respective mobile device;
mapping the location to at least one proximate landmark;
storing the proximate landmarks in association with the identifier; and
discarding the received location;
generating, by the server, a proximate geo model, comprising:
featurizing the proximate landmarks to generate a set of geo model features; and
training a proximate geo model using the geo model features; and
responsive to receiving, by the server and from an ad exchange, a bid request and a target identifier corresponding to an advertising opportunity, and to receiving a consumption history and record of proximate landmarks associated with the target identifier:
applying the proximate geo model to the consumption history and record of proximate landmarks associated with the target identifier to determine a value of the advertising opportunity to an advertising campaign; and
returning, to the ad exchange, a bid amount based on the determined value;
wherein the time passage between receiving the bid request and returning the bid amount is a fraction of a second.
2. The method of claim 1 , wherein a proximate landmark is within a pre-determined radius around the corresponding location.
3. The method of claim 1 , wherein at least one proximate landmark is a type of business.
4. The method of claim 1 , wherein at least one proximate landmark is a type of natural feature.
5. The method of claim 1 , wherein featurizing the proximate landmarks comprises featurizing the proximate landmarks by frequency.
6. The method of claim 1 , wherein featurizing the proximate landmarks comprises:
featurizing the proximate landmarks that are most strongly correlated to converters.
7. The method of claim 1 , wherein featurizing the proximate landmarks comprises:
receiving landmarks of interest to an advertiser; and
featurizing the proximate landmarks that correspond to the landmarks of interest to the advertiser.
8. A non-transitory computer-readable storage medium storing processor-executable computer program instructions that, when executed, cause a computer processor to perform a method, the method comprising:
privatizing, by a server, geographic location information, comprising:
for each of a plurality of mobile devices, for each of a plurality of locations:
receiving a location and an identifier corresponding to a respective mobile device;
mapping the location to at least one proximate landmark;
storing the proximate landmarks in association with the identifier; and
discarding the received location;
generating, by the server, a proximate geo model, comprising:
featurizing the proximate landmarks to generate a set of geo model features; and
training a proximate geo model using the geo model features; and
responsive to receiving, by the server and from an ad exchange, a bid request and a target identifier corresponding to an advertising opportunity, and to receiving a consumption history and record of proximate landmarks associated with the target identifier:
applying the proximate geo model to the consumption history and record of proximate landmarks associated with the target identifier to determine a value of the advertising opportunity to an advertising campaign; and
returning, to the ad exchange, a bid amount based on the determined value;
wherein the time passage between receiving the bid request and returning the bid amount is a fraction of a second.
9. The medium of claim 8 , wherein a proximate landmark is within a pre-determined radius around the corresponding location.
10. The medium of claim 8 , wherein at least one proximate landmark is a type of business.
11. The medium of claim 8 , wherein at least one proximate landmark is a type of natural feature.
12. The medium of claim 8 , wherein featurizing the proximate landmarks comprises featurizing the proximate landmarks by frequency.
13. The medium of claim 8 , wherein featurizing the proximate landmarks comprises:
featurizing the proximate landmarks that are most strongly correlated to converters.
14. The medium of claim 8 , wherein featurizing the proximate landmarks comprises:
receiving landmarks of interest to an advertiser; and
featurizing the proximate landmarks that correspond to the landmarks of interest to the advertiser.
15. A system comprising:
a processor; and
a non-transitory computer-readable storage medium storing processor-executable computer program instructions that, when executed, cause a computer processor to perform a method, the method comprising:
privatizing, by a server, geographic location information, comprising:
for each of a plurality of mobile devices, for each of a plurality of locations:
receiving a location and an identifier corresponding to a respective mobile device;
mapping the location to at least one proximate landmark;
storing the proximate landmarks in association with the identifier; and
discarding the received location;
generating, by the server, a proximate geo model, comprising:
featurizing the proximate landmarks to generate a set of geo model features; and
training a proximate geo model using the geo model features; and
responsive to receiving, by the server and from an ad exchange, a bid request and a target identifier corresponding to an advertising opportunity, and to receiving a consumption history and record of proximate landmarks associated with the target identifier:
applying the proximate geo model to the consumption history and record of proximate landmarks associated with the target identifier to determine a value of the advertising opportunity to an advertising campaign; and
returning, to the ad exchange, a bid amount based on the determined value;
wherein the time passage between receiving the bid request and returning the bid amount is a fraction of a second.
16. The system of claim 15 , wherein a proximate landmark is within a pre-determined radius around the corresponding location.
17. The system of claim 15 , wherein at least one proximate landmark is a type of business.
18. The system of claim 15 , wherein at least one proximate landmark is a type of natural feature.
19. The system of claim 15 , wherein featurizing the proximate landmarks comprises featurizing the proximate landmarks by frequency.
20. The system of claim 15 , wherein featurizing the proximate landmarks comprises:
featurizing the proximate landmarks that are most strongly correlated to converters.
21. The system of claim 15 , wherein featurizing the proximate landmarks comprises:
receiving landmarks of interest to an advertiser; and
featurizing the proximate landmarks that correspond to the landmarks of interest to the advertiser.