Analytics Workflows

Deploy graph intelligence into production. Move from manual exploration to automated, real-time analytics pipelines.

Graph intelligence is only valuable when it is accessible and actionable for your team. Raphtory supports a variety of workflows, from ad-hoc data science to high-scale production monitoring.

Common Workflows

1. Exploratory Analysis

Interactive investigation and hypothesis testing for data scientists and analysts.

• Platform: Jupyter Notebooks or the interactive GraphQL UI.
• Use Case: "Is there a relationship between transaction velocity and account age in our recent fraud cases?"
• Result: Refined temporal motifs that can be promoted to automated alerts.

2. Automated Alerting

Real-time monitoring of live data streams to detect specific temporal patterns as they happen.

• Operation: Raphtory runs continuous windowed queries against an incoming stream (e.g., Kafka).
• Use Case: Alert a security analyst immediately when a "Beaconing" pattern is detected between an internal host and a known C2 server.
• Result: Significant reduction in "Mean Time to Detect" (MTTD).

3. ML Feature Engineering

Using the temporal graph to enrich machine learning models with sophisticated structural features.

• Operation: Generating graph embeddings or node metrics (like temporal PageRank) to use as inputs for traditional ML classifiers.
• Use Case: Feeding network centrality scores into a churn prediction model to account for the "social" nature of user departures.
• Result: 15-25% improvement in model accuracy over non-graph alternatives.

4. Scheduled Reporting

Generating periodic snapshots of network health and risk for executive dashboards.

• Operation: Automated scripts that calculate community health scores or infrastructure bottleneck metrics every 24 hours.
• Use Case: Daily compliance reports showing "High Risk" money flows that crossed international borders.

Deployment Stages