How a U.S. Government Agency is Ensuring Traveler Safety by Imagining What’s Possible with Data

The task of keeping citizens safe in an environment where global security is increasingly a concern is complicated. It requires collecting, processing, querying, and constructing massive amounts of data to screen individuals against risk criteria for threat prevention, public health, and a variety of other use cases. 

Unfortunately, the necessary data is stored across a myriad of sources, from reservation systems to flight manifests and even databases of disease outbreaks. Because the data is siloed and stored in many different formats, it is practically impossible to pre-screen travelers accurately and efficiently.

Closing In On More Efficient Security

But imagine the possibilities if all of this data was presented in a unified view. Think how much easier it would be for government security professionals to ensure that known persons of interest are being identified when they travel.  

One U.S. government agency did just that. They developed and publicly released a turnkey application that gives nation-states and border security entities the necessary data and capabilities to help ensure traveler safety. The application also extends beyond international security into areas such as public health by ensuring that travelers from regions where recent disease outbreaks have occurred receive health screenings. 

Advancing Entity Identification Algorithms

In developing the application, the agency wanted to adopt the most advanced technologies and techniques available. The areas of focus were  data visualization, predictive modeling, and entity resolution—the task of disambiguating records that correspond to real-world entities and detecting relationships among them.

Rapid and scalable entity resolution based on multiple inbound streams of data was the top priority. The application had to answer fundamental questions such as: who is the traveler, and what is his or her profile over time? To answer this and other questions, the agency needed three inputs: derogatory lists, Passenger Name Record (PNR) from reservation system information, and Advanced Passenger Information (API) from airline manifests. 

Adopting Human-Guided Machine Learning

In evaluating options to unify the incoming and ever-changing data and incorporate advanced algorithms to obtain increasingly accurate results, the agency quickly decided against a rules-based Master Data Management (MDM) approach.With any rules-based, MDM approach, it would be nearly impossible to unify incoming source data, scale efficiently, and rapidly improve outputs using advanced matching algorithms. Instead of MDM, the agency supported more advanced entity resolution and matching algorithms in a unique way—human-guided machine learning. 

Making Smarter, Faster Decisions About Travelers

Today, the application provides all of the features necessary to make decisions about travelers. It can receive and store air traveler data, both API and PNR, provide real-time risk assessments against this data based on a country’s own specific risk criteria, and show high-risk travelers as well as their associated flight and reservation information. 

Using the human-guided machine learning approach to government data management, the agency also can enhance basic name/date of birth data and support more advanced entity identification and matching algorithms, including: 

  • Resolving a passenger entity against a known list of people by using the biographic selectors available in the API/PNR data
  • Resolving and matching a person’s identity transmitted via API with one from PNR to create a complete travel instance for the passenger
  • Resolving a passenger entity against the entire population of historical data to build a unique yet comprehensive profile of an individual over time

Maintaining International Security

The application enhances international security, increases agency efficiency, and reduces costs in a variety of ways. One is its ability to employ machine learning to identify potential matches of international traveler records within and across disparate datasets from sources like reservation system and flight manifests. It delivers trusted results because the machine learning itself is guided by input from subject matter experts well-versed in travelers from a particular region. And, by improving entity resolution of travelers’ data from different sources, the solution reduces the inefficiency, inconvenience, and associated costs of unnecessary secondary screenings.

Ultimately, the solution plays a key role in maintaining U.S. and international security and public health. It’s a testament to how seemingly unsolvable problems can be surmounted when organizations imagine what can be accomplished with clean, unified data.

We’d love the opportunity to demonstrate what our customers refer to as the ‘art of the possible’—in person. Schedule a meeting with us here.



Sohaiyla Khalili is the Product Marketing Manager at Tamr. She is responsible for the product content strategy and execution, as well as enablement strategies for sales and field engineering. Before joining Tamr, she worked in product management & product marketing for several technology products and industries. Sohaiyla has a Masters in Engineering Management and a BS in Chemical Engineering.