The National Missing and Unidentified Persons System: Using Search to Connect the Dots

September 2018

Background

Every year in the United States, more than 600,000 people go missing, with as many as 90,000 people missing at any given time. Many of these missing individuals, thankfully, are quickly reunited with their family and loved ones. Unfortunately, thousands of people remain missing for a year or longer. Recognizing the need to improve access to information related to missing and unidentified persons, the National Institute of Justice created the National Missing and Unidentified Persons System (NamUs).

"Users are unable to easily search across both databases at the same time. Finding matches between the Unidentified Persons database and the Missing Persons database becomes time consuming and resource intensive."Since NamUs was created, two separate databases have been launched, which include the NamUs Unidentified Persons (UP) database and the NamUs Missing Persons (MP) database. These databases were developed to provide medical examiners and coroners with a way to store and share case information with law enforcement and family members nationwide. While these databases have provided an incredible tool for those involved in trying to find or identify missing or unidentified people, users are unable to easily search across both databases at the same time. Finding matches between the NamUs UP database and the NamUs MP database becomes time consuming and resource intensive.

 

Using Search to Connect the Dots

The search engine made it easy to create a single, searchable digital catalog of both databases. Machine learning made it easy to not only find matches on distinct values – like eye color, height, date missing, or date found – but also find similarities based on descriptions of a person's physical attributes and circumstances of their case.Seeing this issue and knowing that Voyager Search could help, one of Voyager’s top developers, Trinity Harrison, used Voyager’s off-the-shelf solution to bridge this small, yet critical, gap. By extracting the data from each database on the NamUs website, Harrison was able to easily create a single, searchable digital catalog of both databases. She was then able to put Voyager’s term vectors to work to enable searching for similarity across fields. This allows users to not only find matches on distinct values — like eye color, height, date missing, or date found — but also find similarities based on descriptions of a person’s physical attributes as well as circumstances of their case. She also added spatial information through the use of natural language processing and geotagging, enabling users to narrow down search results spatially as well as using Voyager’s powerful filtering tools.

 

“I’ve been following a few specific cases of missing women on social media for a while. Their stories don’t get any national media coverage and, understandably, their families feel helpless. I observed people trying to help in whatever way they could, including using the NamUs data to find potential matches and bring those families closure,” said Harrison. “A major roadblock is that the missing persons and unidentified persons datasets are separate and there’s no way to do a combined search across both at the same time, or any way to find similarities. After reading comments from people asking for better search tools, I realized Voyager would be perfect to help solve the issue and so I went to work ingesting the data.”

For more information contact:

Voyager Search
Kris Goodfellow, COO - 949-335-2437 - kris@voyagersearch.com
Brian Goldin, CEO - 909-335-2437 - brian@voyagersearch.com

Web Design and Web Development by Buildable