Geolocation Clustering for Device-to-Household Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in accurately associating digital devices with specific households or geographic areas, especially when there is no clear link between devices, and in integrating offline data with device-level information such as digital content consumption and geolocation data, which is often aggregated at less granular levels for privacy or practical reasons.
Innovation Solution
The system employs clustering technology to analyze geolocation data from mobile devices to identify likely household locations by organizing latitude-longitude pairs into clusters, calculating scores based on location frequency and adjacency, and associating devices with geographic areas, including households, neighborhoods, or larger regions, and extends this association to non-mobile devices sharing the same IP address.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If device-level information is collected and stored via IP address or mobile ad ID, then data collection is simplified, but the ability to connect device data to offline household information is lost
Solution Approach 1:
The patent introduces geolocation data as an intermediary element that bridges device-level information (collected via IP address or mobile ad ID) and offline household information. By collecting latitude-longitude pairs from mobile devices and clustering them to identify household locations, the system creates a connecting link that enables association between previously disconnected data ecosystems without changing how device data is initially collected
Solution Approach 2:
The patent segments the association process into distinct components: (1) collecting geolocation data from mobile devices, (2) clustering geolocation points to identify household locations, (3) matching clustered locations with offline household databases, and (4) associating devices with households. This segmentation allows each component to be optimized independently while solving the overall connection problem
2Object-affected harmful factors
If data is aggregated at less granular geographic areas for privacy or practical reasons, then privacy concerns are reduced, but the ability to associate devices with specific households is diminished
Solution Approach 1:
The patent implements a dynamic association system that adapts to data availability. When granular household-level geocoding is available, the system associates devices with specific households. When only aggregated geographic area data is available (for privacy or practical reasons), the system dynamically adjusts to associate devices with broader geographic regions, maintaining functionality across different privacy granularity levels
Solution Approach 2:
The patent changes the granularity parameter of geographic association based on data availability. The system can operate at multiple levels of geographic detail (from specific household addresses to neighborhoods to larger regions), adjusting the association precision parameter according to the quality and privacy requirements of the underlying geocoding data
3Quantity of substance
If home address databases lack granular geocoding, then data availability is improved, but direct device-to-household association cannot be enabled
Solution Approach 1:
The patent applies partial action by implementing a tiered association approach. When granular household-level geocoding is unavailable in certain regions, the system performs association at the next available level of granularity (neighborhood or geographic area). This partial implementation maintains functionality in regions with limited data while achieving optimal precision where granular data is available
Solution Approach 2:
The patent creates a universal association system that functions across diverse data environments. The same geolocation clustering and scoring framework operates whether data is available at household, neighborhood, or regional levels, making the system adaptable to different countries and regions with varying database granularities
Data Source
AI summary
Systems and techniques are disclosed for statistically associating mobile devices and non-mobile devices with geographic areas. One of the methods includes for each selected mobile device of a plurality of mobile devices, receiving latitude-longitude pairs for the selected mobile device, the latitude-longitude pairs corresponding to a location of the selected mobile device during a time period. The plurality of latitude-longitude pairs are organized into clusters, with the clusters corresponding to geographic regions visited by the selected mobile device during the time period. A score is calculated for each cluster, the score representing a likelihood that a user of the selected mobile device resides in a household within the cluster. A location of the user's household is identified to be within one of the geographic areas corresponding to the cluster having the highest score. The mobile device is associated with the geographic area having the highest score.