Anomalías en la contratación pública del Perú: base de ítems adjudicados, indicadores de riesgo y validación externa, 2022–2024

The unit of analysis is the awarded procurement item. The database integrates information on contracting entities, winning suppliers, procurement procedures, number of bidders, reference and awarded amounts, timing variables, supplier–entity relationships, concentration indicators, consortium participation, and traditional procurement risk flags.</p><p dir="ltr">The dataset also includes a traditional risk index based on six red-flag indicators and external validation variables related to subsequent supplier sanctions and contractual penalties.

These validation variables are intended for ex post assessment and should not be used as predictors in unsupervised anomaly-detection models.</p><p dir="ltr">The database was constructed from official public procurement information from Peru’s Organismo Especializado para las Contrataciones Públicas Eficientes (OECE/SEACE) and related official sanction and penalty records. It is designed to support reproducible research using unsupervised machine-learning methods, including Autoencoders and Isolation Forest, for the detection and explanation of atypical procurement patterns.</p><p dir="ltr">The dataset covers 201,524 awarded items from 2022 to 2024 and includes variables related to competition, prices, procedures, timing, supplier–entity relationships, traditional risk indicators, and external validation outcomes.</p>