Skip Navigation
Official website of the United States Government U.S. Department of the Treasury
Bureau of the Fiscal Service Home
Do Not Pay®

Data Analytics

DNP provides free advanced payment analysis service to federal agencies to help combat improper payments.

Analyze payment data for indicators that a payment is being made in error or is vulnerable to abuse

Develop risk scoring to help agencies prioritize and manage reviewing and investigating cross-matches

Screen payees for eligibility such as identifying deceased beneficiaries

A few of the analytics products we offer:

Verification of Payee Eligibility

  • Matching payee data to eligibility data sources

Data Quality Checks

  • Techniques that identify hard-to-detect payment errors and fraud risk in payments, invoices, or vendors

Preventative Controls Development

  • Demonstrate if internal controls are effectively preventing improper payments

Custom Research & Ad-hoc Analysis

  • Provide insight into specific research questions
  • Assist with ad-hoc needs (i.e. cross-government overlap)

Data Availability

  • DNP has access to several datasets to conduct in-depth data analysis

Sample Analysis Techniques

Analytics uses many different techniques to identify a program’s improper payments, such as conducting trend analysis, developing predictive models, and identifying overlapping payments.

  • Helping prioritize resources using risk scoring
  • Conducting pattern and trend analysis on payment data
  • Matching payment data to a death source using advanced and fuzzy matching techniques
  • Detecting anomalies such as same unique identifier (e.g., TIN, SSN, EIN, DUNs) and different names
  • Detecting duplicates (e.g., same payee, same agency, same amount, same date)
  • Identifying data quality issues such as missing or illegitimate TINs
  • Identifying fraud risk such as inconsistent payee information or improper overlap of benefit collection
  • Performing other analysis at the request of the agency
  • Circos Graph -  Ability to turn tables into images

Last modified 08/11/20