CaseLinker: An Open-Source System for Cross-Case Analysis of Internet Crimes Against Children Reports

TL;DR

CaseLinker is an open-source system designed to facilitate cross-case analysis of child sexual exploitation reports by integrating data extraction, visualization, and clustering to identify patterns and support investigations.

Contribution

The paper introduces CaseLinker, a modular hybrid system that combines regex and semantic analysis for extracting, visualizing, and clustering CSEA case data, improving cross-case analysis capabilities.

Findings

Effective extraction of case details from reports

Successful clustering of similar cases using weighted Jaccard similarity

Automated generation of insights and interactive visualizations

Abstract

Child sexual exploitation and abuse (CSEA) case data is inherently disturbing, fragmented across multiple organizations, jurisdictions, and agencies, with varying levels of detail and formatting, making cross-case analysis, pattern identification, and trend detection challenging. This paper presents CaseLinker, a modular system for ingesting, processing, analyzing, and visualizing CSEA case data. CaseLinker employs a hybrid deterministic information extraction approach combining regex-based extraction for structured data (demographics, platforms, evidence) with pattern-based semantic analysis for severity indicators and case topics, ensuring interpretability and auditability. The system extracts relevant case information, populates a comprehensive case schema, creates six interactive visualizations (Timeline, Severity Indicators, Case Visualization, Previous Perpetrator Status, Environment/Platforms, Organizations Involved), provides a platform for deeper automated and manual analysis, groups similar cases using weighted Jaccard similarity across multiple dimensions (platforms, demographics, topics, severity, investigation type), and provides automated triage and insights based on collected case data. CaseLinker is evaluated on 47 cases from publicly available AZICAC reports (2011-2014), demonstrating effective information extraction, case clustering, automated insights generation, and interactive visualization capabilities. CaseLinker addresses critical challenges in case analysis including fragmented data sources, cross-case pattern identification, and the emotional burden of repeatedly processing disturbing case material.