Crisis-Recovery Data Analytics Engine for Volatile Consumer Behavior
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data analytics systems are inadequate for handling volatility in input data during crisis periods, as machine learning models trained on stable historical data struggle to make accurate predictions when faced with inconsistent input data, limiting their ability to forecast demands and optimize supply chains.
Innovation Solution
A crisis-recovery data analytics engine generates a consolidated consumer activity index (CAI) using a crisis-recovery-based machine learning model, which tracks consumer behaviors across pre-crisis, crisis, and recovery periods, incorporating mobility and spending data to provide a quantified score indicating recovery levels, allowing for intelligent data aggregation and visualization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning models are trained on stable historical data, then forecasting accuracy is improved under normal conditions, but the models cannot make accurate predictions during crisis periods with volatile input data
Solution Approach 1:
The system dynamically adapts the machine learning model based on data stability assessment. When volatility is detected in input data, the system transitions from using historically trained models to using crisis-period specific models, allowing the forecasting system to adjust its behavior according to environmental conditions
Solution Approach 2:
The system changes key parameters including the training data period (from historical to crisis-period data), the assessment window for volatility detection, and the forecasting horizon. These parameter adjustments enable the model to operate effectively across different market conditions
2Productivity
If conventional data analytics systems use historically trained machine learning models, then they operate efficiently with stable data, but they fail to identify and use insights from crisis-period data
Solution Approach 1:
The system performs preliminary assessment of input data stability before committing to a forecasting approach. This preliminary action detects volatility patterns and triggers the appropriate model selection and retraining process, ensuring reliable predictions are made only when data quality is sufficient
Solution Approach 2:
The system continuously monitors input data for stability and uses this feedback to determine when to switch between different forecasting approaches. The feedback loop assesses data quality in real-time and adjusts model usage accordingly, maintaining reliability across changing conditions
3Adaptability or versatility
If machine learning models are retrained with crisis-period data, then adaptability to volatile data is improved, but extensive historical data and computational resources are required
Solution Approach 1:
Instead of completely retraining models with all available data, the system uses partial retraining with specifically selected crisis-period data that is most relevant to current conditions. This partial action approach achieves adaptability while consuming fewer computational resources
Solution Approach 2:
The system segments the training data into distinct periods (historical stable data vs. crisis-period data) and uses appropriate segments for different forecasting needs. This segmentation allows efficient use of data resources by applying the right data subset for each forecasting scenario
Data Source
AI summary
Methods, systems, and computer storage media for providing a data analytics index associated with a crisis-recovery data analytics engine in a data analytics system. The data analytics index is a consolidated single index representation of a set of variables associated with a set of consumer behaviors. The crisis-recovery data analytics engine supports generating the data analytics index associated with a pre-crisis period and a crisis-recovery period. In operation, a crisis-recovery dataset—associated with a set of variables of a set of consumer behaviors that support quantifying recovery from a crisis event—is accessed. The set of consumer behaviors are selected based on a crisis-recovery machine learning model that is trained on a pre-crisis dataset and a crisis dataset for selecting the set of consumer behaviors. A data analytics index is generated based on the crisis-recovery dataset. A data visualization comprising crisis-recovery data indicating recovery from the crisis event is generated.