Quarterly Journal of Emerging Technologies and Innovations

Data-Driven Perspectives on Federal Budgetary Dynamics for Identifying Anomalies and Patterns in Resource Allocation and Obligation Trends

Rafiqus Salehin Khan

Data & Assessment Associate at Success Academy Charter Schools, NYC

https://orcid.org/0009-0002-5634-4681

Md Riad Mahamud Sirazy

Technology Integration Associate (Information Technology), Shine Electronics Co. NYC

https://orcid.org/0009-0009-4838-1635

Rahul Das

Device Quality Engineer, Shine Electronics Co. NYC

https://orcid.org/0009-0008-0365-4399

Sharifur Rahman

MSc in Business Analytics, University of Central Oklahoma, USA

https://orcid.org/0009-0009-5465-7159

Keywords: anomalies, budgeting practices, clustering models, efficiency, resource allocation, spending patterns, outlier detection

---

Abstract

Federal agencies tasked with significant program mandates often exhibit uneven spending patterns that can obscure broader fiscal objectives. This study investigated how major sub-components allocate and obligate resources over multiple years, seeking to identify common behaviors, anomalies, and possible influences on spending decisions. Objectives included examining efficiency, detecting outliers, and grouping sub-components according to their historical trends to guide refined budgeting practices. Methods encompassed K-means and hierarchical clustering, correlation analysis to assess alignment between resources and obligations, outlier detection through z-score and interquartile range, and linear regression to quantify directionality in spending patterns. Results highlighted several trajectories, with the “Centers for Medicare and Medicaid Services” emerging as a clear outlier due to rapid escalation in obligations that exceeded standard expectations. The “Office of Federal Student Aid” showed a contrasting scenario, underspending its allocated resources and displaying negative z-scores by 2023. Meanwhile, the “Food and Nutrition Service” maintained a balanced ratio of obligations to allocations, pointing to stable management processes. Clustering models grouped agencies into stable, quickly increasing, and fluctuating spending categories, revealing operational parallels among sub-components that share cluster assignments. Correlation tests affirmed a strong linkage between resource distribution and spending outcomes in most cases, notwithstanding a few anomalies. Recommendations highlight the importance of embedding advanced data analytics in budgetary supervision, reinforcing real-time monitoring to enable early detection of deviations and fostering inter-agency collaboration in the dissemination of best practices. Forward-looking strategies focused on adaptive budget allocations and methodological refinements promise stronger alignment between designated resources and evolving policy imperatives.