Online System Predicting Location Traffic Deviations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Publishing users on online systems often miss opportunities to present content to a higher number of users within a threshold distance of their location due to unawareness of events that increase user numbers, leading to decreased exposure and revenue.
Innovation Solution
The online system analyzes event data and user location information to identify deviations in user numbers relative to historical averages, sending notifications to publishing users when a significant increase in users is anticipated, allowing them to optimize content presentation and advertising strategies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If publishing users determine content presentation amounts based on historical user numbers, then content presentation costs are controlled, but opportunities to present content during peak user periods are missed
Solution Approach 1:
The system performs preliminary analysis of event data and user location information before publishing users make content presentation decisions. By predicting peak user periods in advance based on scheduled events and historical patterns, the system enables publishing users to optimize their content presentation timing and maximize exposure during high-traffic periods.
2Productivity
If publishing users increase content presentation during peak periods, then user exposure and revenue increase, but system complexity for tracking and analyzing user data increases
Solution Approach 1:
The online system leverages its existing multi-functional capabilities by repurposing its event management infrastructure and user location tracking (already used for other purposes) to also predict peak user periods for content presentation optimization. This avoids adding significant complexity while enabling new revenue-optimizing functionality.
Solution Approach 2:
The system implements feedback loops by continuously monitoring actual user presence at locations during and after events, comparing predicted peak periods with actual usage patterns. This feedback refines future predictions and allows the system to automatically adjust content presentation recommendations to publishing users based on validated insights.
3Measurement precision
If the online system tracks user locations and event data in real-time, then accurate predictions of peak user periods are achieved, but data processing requirements and computational resources increase
Solution Approach 1:
The system performs preliminary processing and aggregation of user location and event data before predictions are needed. By pre-processing this data and establishing baseline patterns in advance, the system reduces the computational burden during real-time prediction operations, achieving accurate forecasts with lower instantaneous processing requirements.
Solution Approach 2:
The system implements selective tracking by focusing computational resources on analyzing user data for locations and time periods where events are scheduled or where historical patterns indicate potential peak periods. Rather than continuously processing all user location data everywhere, the system applies partial action only where and when predictions are most valuable.
Data Source
AI summary
An online system predicts whether a location will experience a threshold increase in traffic over the location's historical average amount of traffic. To predict a future deviation over historical traffic, the online system identifies events within a threshold distance of the location and determines an average number of indications that users will attend events within a threshold radius of the location during a prior time interval. The online system determines a total number of indications that users will attend future events within the threshold distance of the location, disregarding locations associated with less than a threshold number of future events and future events for which the online system received less than a threshold number of indications that users will attend, and determines a ratio of the total number of indications to the average number of indications that users will attend received for the prior events during the time interval.