ICE seeks support for analytics platform
Immigration and Customs Enforcement’s Homeland Security Investigations unit is holding an April 28 industry day to discuss the program and acquisition status of its Repository for Analytics in a Virtualized Environment.
Since 2018, the RAVEN team has been has been working on its cloud-based next-generation analytics platform that will give HSI users in law enforcement a suite of search, analysis and reporting tools to run across raw or unevaluated datasets. Now, HSI is preparing to issue contracts for three support services categories: data analytics, user interface/user experience and DevSecOps.
RAVEN’s capabilities come from tools developed or purchased by HSI’s eight divisions and 185 field offices in an effort to reduce duplication and use existing tools for multiple tasks. The program also aims to combine and maximize the efficiency and capabilities of open-source tools, according to a May 2020 privacy impact assessment
Rather than replace ICE’s existing case management systems, RAVEN will primarily be used for large, complex analytical projects -- curating and linking “seemingly disparate raw datasets by performing advanced analytics across multiple datasets, thus enabling users to accomplish tasks currently considered too large or complex for existing systems,” the PIA said.
In an HSI trial of RAVEN’s advanced analytics capabilities, the system reduced the review of subpoenaed telephone records from over 50 hours to just 30 seconds. It was able to ingest and process 16 different worksite enforcement audits, analyzing 55,278 individual I-9s and saving 289 working days or 2,312 hours. In addition, RAVEN used natural language processing to extract information that aided in identifying 760 children as outbound air travelers who may be linked to family unit fraud or child recycling, according to the 2021 ICE budget overview.
RAVEN has five primary functions, according to an HSI presentation accompanying the April 14 notice:
- Automated data collection of internet and non-public online information to support criminal investigations and investigative initiatives.
- Automated extraction, ingestion, transformation and integration of information, including connections to data in existing Department of Homeland Security systems.
- Intelligent analysis that fuses information from different sources to create a single threat picture and enables uses to search across all available data to expose complex event interactions and criminal networks.
- Collaboration and sharing services that allow users from different agencies to work together securely; a fully integrated knowledge transfer system that supports sharing of information on constantly evolving criminal typologies; and intelligent analytics to test, support or refute assertions and expose non-obvious trends.
- User-friendly visualization and reporting tools that can be easily learned and accessed via web and mobile devices and exported for sharing with interested partners.
According to the HSI presentation, RAVEN is designed as a hub and spoke model. It features a single shared data layer, a reusable microservice for data cleaning and machine learning – all surrounded by a suite of special-purpose web and mobile user interfaces.
HSI intends the platform to be modular and extendable so investigators can keep pace with evolving criminal networks by rapidly building new analytics without changing the underlying framework. Ease of use is critical. The platform must be easy to use for both general and expert users, and accessible through the web and a custom app.
Read more about the RAVEN project and the upcoming industry day here.
Connect with the GCN staff on Twitter @GCNtech.