Contextual data
Enrich real-time streams with contextual data
Overview
For Joule applications, contextual data is crucial for enabling advanced and insightful stream processing. By seamlessly integrating contextual data with real-time events, the system delivers enriched processing outcomes and better informed insights.
Within Joule, contextual data comprises both static and dynamic data:
Static data includes reference information, such as customer contracts, product SKUs, or start-of-day FX rates, etc;.
Dynamic data consists of real-time metrics that drive KPI computations, machine learning (ML) feature engineering, analytics, and other performance-based measures.
This integration allows Joule to deliver impactful insights through capabilities such as data transformation, predictive analytics and real-time KPI computations.
Some example applications include:
Geospatial Location Analytics: Leveraging well-known areas of interest for real-time geospatial insights.
Real-Time Customer Promotions: Tailored promotions based on historical buying profiles and real-time customer behaviour.
Multi-Currency Dynamic Pricing: Adapting pricing strategies based on live exchange rates and market conditions.
Real-Time Spending Alerts: Powered by pre-engineered ML features to track and alert users of unusual spending patterns.
Predictive Capacity Demand Forecasting: Using live metrics to forecast future resource or capacity needs with high accuracy.
These capabilities enable Joule to not only process data in real time but also drive actionable insights that influence business decisions.
Key features of Joule's contextual data capability
Event enrichment Adds static information (i.e., geographic or device data) to live events for richer analysis.
Transformation and validation Transforms and validates event attributes using predefined patterns or rules.
Real-time predictive analytics Utilises contextual data such as machine learning models for predictions on live data.
KPI computation Generates real-time performance indicators based on enriched event data.
Local caching Contextual data is treated as "slow-moving" and often cached locally within processing nodes to optimise performance.
Example use case
To understand localised network performance, a telecommunications (telco) provider can leverage contextual data by enriching mobile telemetry events with:
Mobile manufacture
Mobile model
and mapping connected cell towers to postal areas
This approach enables the telco to assess network performance in specific areas, identifying regions with strong or weak signals and pointing towards where to improve customer experience.
Joule's contextual data in action
In machine learning-driven use cases, this process enables real-time predictions with static model references, as contextual data can be stored in local caches to optimise speed and efficiency.
Joule provides processors with implementation interfaces to access and apply this cached data efficiently within each processing context. Common configurations include specifying data sources, caching strategies and defining how and when contextual data is refreshed.
Contextual data is treated differently to event driven data, it is considered slow-moving
By providing localised, cached, data stores within the processing context. Joule provides processors the required implementation interfaces to access and apply this data within a localised stream processing context.
Architecture
Understand how contextual data integrates within Joule
Configuration
Set up a contextual data-driven use case
Types of source implementations
Joule integrates with various data sources to support contextual data within stream processing, these are:
Last updated