Executive KPI Summary — Power BI Layer
Total crossings — all time
11.5B
273,391 records processed · 114 ports
US-Mexico crossings (2025)
266.2M
▲ 0.06% vs 2024 — stable corridor
US-Canada crossings (2025)
60.3M
▼ 19.2% vs 2024 — anomaly flagged
Anomaly ports detected (z > 2.0)
10
KMeans + z-score methodology
Monthly crossing volume trend
2019–2025 · US-Mexico vs US-Canada Border · millions of crossings
US-Mexico Border
US-Canada Border
Crossing type breakdown
2020–2025 · share of total crossings by mode
Top 15 ports by total volume
2020–2025 · millions of crossings
Year-over-year comparison by border
2023 vs 2024 vs 2025 · millions
Analyst Drill-Down — Tableau Layer
Port volume distribution — top 15 ports
Orange = US-Mexico · Blue = US-Canada · millions of crossings (2020–2025)
Crossing mode distribution
All borders · 2020–2025
Monthly volume — COVID impact & recovery
US-Mexico border · 2019–2023 · millions
Anomaly-flagged ports
Z-score outlier detection · threshold z > 2.0 · Python scikit-learn · 2020–2025 monthly data
| # | Port | State / Border | Anomaly events | Max z-score | Signal strength | Status |
|---|
Port Risk Clustering — Big Data + ML Layer
Port risk tier distribution
KMeans k=4 · 114 active ports · features: volume, volatility, measure types
Critical Hub — 1 port
High Volume — 11 ports
Moderate — 45 ports
Low Activity — 57 ports
Volume vs volatility scatter
Each point = 1 port · color = risk tier · x = total volume · y = volatility index
Port risk register — full classification
Top 20 ports by volume · KMeans cluster assignment · Scikit-learn StandardScaler normalization
| Port | State | Border | Total volume (2019–2025) | Volatility index | Risk tier | Resource priority |
|---|
Business Analysis Findings Memo — Deliverable Layer
Business Analysis Findings Memo
Subject: CBP Border Crossing Intelligence — Pattern Analysis & Resource Recommendations
Prepared for: GDIT Intelligence & Homeland Security Division · Springfield, VA
Date: April 2026 · Dataset: Bureau of Transportation Statistics · data.gov
Prepared for: GDIT Intelligence & Homeland Security Division · Springfield, VA
Date: April 2026 · Dataset: Bureau of Transportation Statistics · data.gov
Finding 01
San Ysidro is the sole Critical Hub — 274M crossings, 2× nearest competitor
KMeans clustering across 114 active border ports (features: total volume, avg monthly, volatility index, measure type diversity) isolated San Ysidro, CA as the only port classified "Critical Hub." It processed 274.4M crossings (2020–2025) — nearly 2× El Paso's 147.2M. Its volatility index of 0.995 indicates dense, consistent throughput with limited seasonal variance. Any disruption at this single node creates cascading effects across the entire southwest corridor. Current resource models that do not account for this concentration underestimate systemic operational risk.
Finding 02
US-Canada border volume dropped 19.2% in 2025 — a statistically significant anomaly
US-Canada crossings fell from 74.6M (2024) to 60.3M (2025) — a 19.2% YoY decline that does not appear in US-Mexico data (which held flat at ~266M). This divergence is statistically significant and warrants root-cause investigation. Possible explanations include seasonal data incompleteness for late 2025, policy-driven traffic suppression, or a genuine behavioral shift. Since Personal Vehicle Passengers account for 54% of all crossings, a sustained decline in leisure and commuter travel would directly affect CBP staffing models and technology deployment decisions at northern ports.
Finding 03
10 ports flagged with anomalous monthly spikes — concentrated in Jul–Aug 2024
Z-score analysis across all port-month combinations (2020–2025) flagged 10 ports exceeding z > 2.0, with Ferry (z=3.73), Columbus (z=3.59), and Lancaster (z=3.42) showing the strongest signals. Critically, 8 of the 10 flagged anomalies cluster in July–August 2024 on the US-Canada border — a temporal concentration that exceeds normal summer seasonality and suggests a coordinated or policy-driven traffic pattern worth monitoring at the operational intelligence level.
Risk 01
Single-point dependency at San Ysidro creates systemic vulnerability
With 274M crossings concentrated at one port — nearly 19% of all US-Mexico volume — any operational disruption creates a cascading backlog across the entire southwest corridor. Current contingency nodes (Otay Mesa: 62M, Calexico: 78M) have insufficient combined capacity to absorb a San Ysidro outage. Recommendation: model alternate routing capacity and pre-position surge staffing protocols at Otay Mesa and Calexico as defined contingency nodes.
Risk 02
Canada volume decline may trigger premature staffing reductions at northern ports
A 19.2% volume decline, if misinterpreted as reduced structural demand rather than a data or policy artifact, could justify staffing drawdowns at northern border ports. If volumes recover in 2026, those ports would face a capacity gap during peak summer months — the exact period when anomalies were already detected in 2024. Recommendation: implement a rolling 24-month staffing model across all ports rather than single-year snapshots.
Recommendation
Implement a tiered resource allocation model across all 4 port risk clusters
Based on KMeans clustering of all 114 active ports into 4 risk tiers, GDIT should propose the following tiered deployment model to DHS/CBP clients:
Tier 1 — Critical Hub (1 port: San Ysidro)
Permanent AI-assisted anomaly monitoring · Dedicated analyst team · Real-time dashboard with executive alerts · Contingency routing protocols
Tier 2 — High Volume (11 ports: El Paso, Laredo, Hidalgo, Calexico, etc.)
Automated anomaly alerts (z-score threshold: 2.0) · Monthly review cycles · Predictive staffing models
Tier 3 — Moderate (45 ports)
Quarterly review · Surge capacity on-call · Shared analyst coverage
Tier 4 — Low Activity (57 ports)
Annual review · Automated reporting only · Flag for consolidation analysis
This framework directly aligns with GDIT's Predictive Tools, Common Operating Picture, and AI/Data Analytics capabilities outlined in the Intelligence & Homeland Security division strategy — and can be operationalized using existing GDIT edge computing and MotionGEOINT infrastructure.
Tier 1 — Critical Hub (1 port: San Ysidro)
Permanent AI-assisted anomaly monitoring · Dedicated analyst team · Real-time dashboard with executive alerts · Contingency routing protocols
Tier 2 — High Volume (11 ports: El Paso, Laredo, Hidalgo, Calexico, etc.)
Automated anomaly alerts (z-score threshold: 2.0) · Monthly review cycles · Predictive staffing models
Tier 3 — Moderate (45 ports)
Quarterly review · Surge capacity on-call · Shared analyst coverage
Tier 4 — Low Activity (57 ports)
Annual review · Automated reporting only · Flag for consolidation analysis
This framework directly aligns with GDIT's Predictive Tools, Common Operating Picture, and AI/Data Analytics capabilities outlined in the Intelligence & Homeland Security division strategy — and can be operationalized using existing GDIT edge computing and MotionGEOINT infrastructure.