The existing system being used by the Law Enforcement Agency was painfully slow to ingest and incorporate new datasets into the environment. The users had been incredibly frustrated as a large number of feeds due to be ingested was due to be a multi-year project. The interface was old and unintuitive leading to a lack of enthusiasm from users, particularly new recruits. The search capability was particularly poor leading to frustration and a lack of user engagement. On top of this the costs were incredibly high due to proprietary hardware and very high software costs.
Siren undertook an initial phase of the project to solve the issues of speed of data ingestion and usability. In the space of 6 weeks Siren stood up a system which had ingested over 100m records from 12 different systems (RMS, CAD, warrants, complaints, arrests etc.), processing that data through Entity Resolution and NLP. The users got hands-on experience with the system straight away. Siren was able to leverage a growing investment in Elasticsearch with a simple install onto an existing cluster. The cluster was then moved to a cloud deployment.
Siren has been enthusiastically adopted by the user community at the law enforcement agency. The search based approach, integrated Entity Resolution and NLP, dashboards and intuitive graph analytics are a game changer. The second phase of the project got the number of feeds up to 45 with more to follow. With so many feeds it was never possible to create a single identifier for “people” so the Entity Resolution capability in Siren creates a runtime linkage of all the records recommended to make up that “person”. This is a massive automation gain, taking over 12 months off of the planned ingestion pipeline. The environment is primed to ingest new data in structured and unstructured form. An example of this was the integration of video from body worn cameras into the Siren environment allowing it to be integrated into the overall investigative environment.