EmmanuelJDP/dgcp-procurement-risk-do

LexMetrix AltData: Dominican Procurement Risk Intelligence (2021–2026)

Public procurement dataset + risk scoring model for monopoly detection, compliance auditing, and open-data research across Latin America.

---

📊 Dataset Overview

Metric	Value
Records	305,094 government contracts
Period	January 2021 – June 2026
Government spending analyzed	USD 14.6 billion
CRITICAL risk flags	2,632 contracts (0.86%)
5+ year monopoly patterns	193 contracts
Single-Bidder Rate	73% (vs. LATAM avg ~18%)
Format	OCDS-compliant JSON + CSV
Source	Dirección General de Contrataciones Públicas (DGCP), Dominican Republic

---

🎯 What This Is

A public intelligence feed of Dominican government contracts with machine-scored risk indicators for:

• Monopoly/cartel detection: Single-bidder awards, HHI concentration indices, temporal clustering
• Compliance auditing: Formal defect patterns, deadline anomalies, procedure irregularities
• Regulatory research: Procurement system benchmarking, policy impact analysis
• Commercial intelligence: Market structure analysis for suppliers, competitive intelligence

Non-purpose: This is not for predicting individual firm behavior or singling out contractors. It's for systemic analysis of procurement patterns.

---

🔍 Key Findings (Dominican Republic, 2021–2026)

Single-Bidder Dominance

• 73% of awards received only 1 bid (vs. healthy baseline ~15–20%)
• Suggests: barriers to entry, collusion, or structural market failure
• 2,632 contracts flagged as CRITICAL for monopoly risk

Temporal Monopoly Patterns

• 193 contracts show continuous single-supplier awards over 5+ years to same firm
• Typical indicators:
  • Same OCID pattern (supplier ID)
  • Consistent within agency
  • No re-tendering or competitive process
  • Works, services, or supplies categories (not construction)

Spending Concentration

• Top 10 suppliers: 12.3% of total award value
• Top 50 suppliers: 31.2% of total award value
• Suggests oligopolistic procurement ecosystem

---

📦 Data Schema

Core Fields (OCDS Standard)

{
  "ocid": "ocds-xxxxx-xxxxx",
  "tender_id": "...",
  "supplier_id": "...",
  "supplier_name": "...",
  "award_date": "2023-05-15",
  "award_value": 1500000,
  "currency": "DOP",
  "tender_title": "Suministro de combustible",
  "buyer_name": "Instituto Nacional de Agua Potable y Alcantarillado",
  "buyer_code": "15627",
  "procedure_type": "direct_award|limited_tender|open_tender",
  "goods_classification": "31820000",  // CPV code
  "contract_status": "active|terminated|completed"
}

LexMetrix Risk Scoring Fields

{
  "single_bidder_flag": true,
  "hhi_index": 10000,  // 0–10000; >2500 = high concentration
  "temporal_monopoly_years": 5.2,  // continuous awards to same supplier
  "temporal_monopoly_flag": true,  // if ≥5 years
  "procedure_irregularity_score": 0.78,  // 0–1; based on OCDS completion
  "risk_category": "CRITICAL|HIGH|MEDIUM|LOW",
  "risk_score": 0.92,  // composite 0–1
  "flagged_reason": "Single bidder + 5yr monopoly pattern",
  "analysis_date": "2026-06-02"
}

---

💡 Use Cases

1. Compliance & Anti-Monopoly Research

• Identify procurement patterns violating competition law
• Screen for re-tendering violations, cartel indicators
• Train machine learning models for fraud detection

2. Regulatory & Policy Analysis

• Benchmark Dominican procurement against LATAM peers
• Measure transparency/competition impact of policy changes
• Analyze spending by agency, procedure type, goods category

3. Commercial Intelligence

• Suppliers: analyze government market structure, identify opportunities
• Investors: assess procurement-driven sector risk
• Media/Academics: investigate systemic inefficiencies

4. Replicability Across LATAM

• Methodology applies to: Colombia (SECOP), Peru (SEACE), Paraguay (DNCP), Chile (ChileCompra)
• HHI calculation, single-bidder detection, temporal clustering = transferable
• OCDS-compliant data from 25+ countries available through Open Contracting Partnership

---

🔧 Methodology Snapshot

Risk Scoring Formula

risk_score = (w1 × single_bidder_norm) 
           + (w2 × hhi_norm) 
           + (w3 × temporal_monopoly_norm)
           + (w4 × procedure_irregularity)

• Weights: w1=0.35, w2=0.30, w3=0.25, w4=0.10 (trained on 50 validated cassation appeals)
• Normalization: min-max scaling within each component
• Threshold for CRITICAL: risk_score ≥ 0.85

Data Quality

• Source: DGCP Portal de Datos Abiertos (official OCDS publication)
• Validation: SHA256 hash tracking, date range checks, supplier ID consistency
• Missing data: <2% across core fields
• Completeness: 98.4% of records have sufficient fields for risk scoring

---

📥 How to Use This Dataset

Python Example

import pandas as pd
import json

# Load dataset
df = pd.read_csv('dgcp_procurement_risk_2021_2026.csv')

# Filter to CRITICAL risk
critical = df[df['risk_category'] == 'CRITICAL']
print(f"CRITICAL contracts: {len(critical)}")
print(f"Total spending at risk: ${critical['award_value'].sum():,.0f}")

# Temporal monopoly analysis
monopoly = df[df['temporal_monopoly_flag'] == True]
print(f"Contracts with 5+ year single-supplier patterns: {len(monopoly)}")

# Single-bidder rate by agency
sbr_by_agency = df.groupby('buyer_name').apply(
    lambda x: (x['single_bidder_flag'].sum() / len(x))
)
print(sbr_by_agency.sort_values(ascending=False).head(10))

SQL Query (if loaded into database)

SELECT 
    buyer_name, 
    COUNT(*) as contract_count,
    SUM(award_value) as total_value,
    ROUND(100.0 * SUM(CASE WHEN single_bidder_flag THEN 1 ELSE 0 END) / COUNT(*), 2) as sbr_pct,
    COUNT(CASE WHEN risk_category = 'CRITICAL' THEN 1 END) as critical_count
FROM contracts
WHERE award_date >= '2021-01-01'
GROUP BY buyer_name
ORDER BY critical_count DESC;

---

🌍 Replicability: Applying to Other LATAM Countries

This methodology is FULLY TRANSFERABLE to other LATAM procurement systems:

✅ COLOMBIA (SECOP) • Single-bidder rate: 45-60% (expected) • HHI + temporal analysis: Same approach • Data: https://www.secop.gov.co/ (OCDS feed)

✅ PERU (SEACE)
• Temporal monopoly detection: Fully applicable • Expected SBR: 40-55% • Data: https://www.seace.gob.pe/ (OCDS API)

✅ PARAGUAY (DNCP) • Risk scoring model: Transferable with recalibration • Similar monopoly patterns expected • Data: https://www.dncp.gov.py/

✅ CHILE (ChileCompra) • Procedure irregularity scoring: Applicable + enhanced precision • Expected SBR: Lower (~35%) due to mature market • Data: https://www.chilecompra.cl/ (OCDS registry)

For replication: Request dataset access to those countries' OCDS feeds through Open Contracting Partnership.

---

📊 Access Full Historical Dataset

This sample (50 rows) is from the complete LexMetrix AltData dataset:

📦 LATAM Procurement Risk Index (2021–2026)

• 305K+ contracts analyzed
• 26,785 market segments (buyer × commodity)
• 6 years of continuous data
• USD 859.6B spending analyzed
• 27 metrics per record (HHI, SBR, temporal, risk scores)

→ Get the Full Dataset (Parquet + CSV + JSONL)

USD 299.99 | One-time purchase | Instant download | Full documentation included

Use cases: ESG compliance, hedge fund due diligence, risk analysis, academic research.

---

---

📋 Citation

If you use this dataset in research, policy analysis, or reporting:

BibTeX

@dataset{lexmetrix_altdata_2026,
  title={LexMetrix AltData: Dominican Procurement Risk Intelligence (2021–2026)},
  author={Emmanuel J. D. P.},
  year={2026},
  publisher={Hugging Face Datasets},
  url={https://huggingface.co/datasets/EmmanuelJDP/dgcp-procurement-risk-do},
  note={305,094 OCDS-compliant government contracts with HHI, single-bidder, and temporal monopoly risk scoring}
}

APA

Emmanuel, J. D. P. (2026). LexMetrix AltData: Dominican procurement risk intelligence (2021–2026) [Dataset]. Hugging Face Datasets. https://huggingface.co/datasets/EmmanuelJDP/dgcp-procurement-risk-do

---

🤝 Contributing & Questions

Have questions about the data?

• Open an issue in the Discussions tab
• Include: country context, use case, expected sample size

Want to extend this dataset?

• Contributions welcome: add risk scoring variants, sector-specific analysis, time-series forecasting
• Contact: [your email / message]

Found an error?

• Report in Discussions with: contract OCID, issue type, suggested correction
• Corrections tracked with version tag

---

📚 Related Datasets & Resources

• Open Contracting Partnership Data Registry — OCDS data from 25+ countries
• AgentPublic/open_government — Multi-country OCDS datasets
• Dominican Republic DGCP Portal — Official source; updates daily
• Open Government Partnership — Policy context, implementation guides

---

📄 License & Terms

Dataset License: CC BY 4.0 (Creative Commons Attribution)

Terms of Use:

• ✅ Use for research, policy analysis, journalism, compliance auditing
• ✅ Cite this dataset if published
• ✅ Share findings with Open Contracting Partnership community
• ❌ Do not use to identify or harass individual suppliers or officials
• ❌ Do not claim authorship or resell without attribution

Data Source: Dominican Republic DGCP (Dirección General de Contrataciones Públicas), published under Open Government Partnership commitments.

---

⚙️ Technical Details

• Last Updated: 2026-06-02
• Update Frequency: Weekly (synced to DGCP daily feed)
• File Formats: JSON Lines (.jsonl), CSV, Parquet
• Total Size: ~850 MB (uncompressed)
• Python Version Required: 3.8+
• Dependencies: pandas, numpy (for analysis)

---

🔐 Reproducibility & Versioning

All analysis scripts are version-controlled with SHA256 hashing:

• Dataset hash: [will be auto-generated by HF]
• Scoring model checksum: a3f8e2c1d9b4... (available on request)
• Validation set: 50 manually reviewed cassation appeals (Ley 2-23, Dominican Republic)
• Accuracy: 93% on validation set (CRITICAL detection)

---

Start exploring: Load the CSV, filter to risk_category == 'CRITICAL', and investigate patterns by agency or procedure type. Questions? Open an issue.