On-Device Location History Hashing for Private Content Attribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to effectively utilize on-device location history for content analytics while ensuring user privacy and preventing unnecessary sharing of location data off the user device.
Innovation Solution
A system and method that processes user location history on the client device, generating de-identified content-location event pairs using hash functions, and transmits only relevant data to a server computing system, employing an attribution pipeline with machine learned models to determine conversions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If location history data is processed on the server, then analytics accuracy is improved, but user privacy is compromised and bandwidth usage increases
Solution Approach 1:
The patent segments the analytics processing into two parts: de-identification and hashing of location data is performed on-device (client segmentation), while only the hashed, de-identified data is transmitted to the server for analytics processing. This segmentation allows analytics to be performed on server while preserving user privacy by removing personally identifiable information before transmission.
Solution Approach 2:
The patent introduces hash functions as an intermediary mechanism that transforms location data into de-identified representations. The hashing process acts as a mediator between raw location data and analytics processing, enabling the server to perform analytics on location patterns without accessing actual user location information, thus resolving the privacy-accuracy contradiction.
2Loss of information
If all location data is transmitted to the server, then analytics completeness is improved, but bandwidth usage and processing time increase
Solution Approach 1:
The patent extracts only the essential analytics-relevant features from location data by performing de-identification and hashing on-device. This extraction process removes personally identifiable information while preserving location pattern data, allowing the server to receive a condensed, privacy-preserving representation that maintains analytics completeness without transmitting unnecessary data volume.
Solution Approach 2:
The patent performs preliminary processing (de-identification and hashing) of location data on the client device before transmission. This preliminary action filters and condenses the data in advance, ensuring that only the necessary de-identified information is transmitted to the server, thereby reducing bandwidth usage while maintaining the completeness needed for accurate analytics.
3Productivity
If identifiable location data is shared with third parties, then advertising effectiveness is improved, but user privacy and security are compromised
Solution Approach 1:
The patent changes the parameter of location data from identifiable to de-identified through hashing transformations. By applying hash functions to location data, the system transforms the data into a form that preserves advertising effectiveness (by maintaining location patterns and context) while fundamentally altering the privacy parameter to remove personally identifiable information, thus resolving the effectiveness-privacy contradiction.
Data Source
AI summary
Example embodiments of the present disclosure provide for an example method for on-device location history for content analytics. The example method includes obtaining data associated with a first user and a second user. The example method includes generating content event groupings associated with the users. The example method includes a user device requesting the content event groupings. The example method includes obtaining de-identified content-location event pairs generated on the user device. The example method includes inputting the de-identified content-location event pairs into an attribution pipeline. The example method includes obtaining attribution data associated with conversions attributed to specific content-location event pairs.