Cristian David Correa-Álvarez

This study compares 4G and 5G mobile network performance across Colombian municipalities using regulatory crowdsourced data. The dataset contains 471,800 indicator records from June 2023 through June 2026. Download speed, upload speed, latency, jitter, and packet loss were matched by municipality, operator, and reporting period. Because published values are rolling three-month means, the primary analysis uses March and June 2026, yielding 1461 complete paired observations across 634 municipalities. Geometric mean ratios with equal municipality weights and municipality-block bootstrap intervals summarize the comparisons. The 5G/4G download ratio is 1.94 (95% bootstrap interval: 1.80–2.09); the corresponding ratios are 1.19 for upload, 0.90 for latency, 0.78 for jitter, and 0.68 for packet loss. All five indicators strictly favor 5G in 38.5% of municipality-weighted observations; this share is 25.9% when each indicator must improve by at least 10%. The download ratio rises to 2.59 with equal observation weights and to 5.49 in the 69 municipalities with complete reporting for all three 5G operators at both periods. Thus, average differences favor 5G, but their magnitude varies with sample composition and weighting, and improvements are not uniform across indicators. The findings describe a first-half-2026 performance snapshot, not causal deployment effects or long-term evolution; second-quarter data remain preliminary.

Este artículo presenta una herramienta metodológica diseñada para el cálculo e interpretación de coeficientes de correlación no paramétricos, orientada a profesionales y estudiantes sin formación estadística formal que requieren analizar asociaciones entre variables en contextos donde no se cumplen los supuestos de normalidad. La propuesta abarca la descripción detallada de los principales coeficientes no paramétricos utilizados en estadística aplicada (Spearman, Kendall, Gamma, Lambda, C de Cramér, Phi y Somers D), incluyendo sus fundamentos teóricos, criterios de aplicabilidad y las condiciones que determinan su uso adecuado. Como parte del desarrollo metodológico, se implementaron estos coeficientes en el lenguaje de programación R, acompañados de ejemplos prácticos con su respectivo código y la creación de una aplicación interactiva en Shiny. Esta aplicación permite al usuario cargar archivos, seleccionar variables y visualizar resultados sin necesidad de conocimientos en programación, lo que amplía su accesibilidad en entornos multidisciplinarios. Para validar la herramienta, se llevaron a cabo estudios de caso en distintos campos, entre ellos un análisis en ingeniería de materiales, en el que se identificó una relación inversa entre la dureza y la resistencia al desgaste en piezas metálicas sometidas a tratamiento térmico. Los resultados obtenidos evidencian que la elección del coeficiente de correlación incide de manera significativa en las conclusiones del análisis, especialmente cuando se trabaja con datos no normales o escalas categóricas. Se concluye que el enfoque no paramétrico constituye una alternativa robusta para el análisis de asociaciones, y que la herramienta propuesta facilita su aplicación en investigaciones técnicas, educativas y científicas que requieren rigor metodológico sin complejidad computacional.

Environmental degradation in Colombia is often studied through separate sectoral lenses, although air pollution, emerging water contaminants, urban noise, mining-related soil contamination, artisanal-mining exposure, and material throughput arise from overlapping institutional and economic conditions. This review compares these domains to identify how environmental-health protection, social equity, resource use, and implementation capacity can inform sustainability governance. A structured integrative synthesis of 62 peer-reviewed, institutional, and conceptual sources organized evidence into cross-domain matrices and six ordinal dimensions: structural entrenchment, human-health burden, ecological propagation, persistence, financing attribution, and implementation complexity. A pattern was treated as convergent when it recurred in at least four domains and retained its meaning under a three-domain sensitivity threshold. Four convergences were identified: lock-in in mobility, production, and livelihood systems; coupled health–ecosystem effects; disproportionate exposure under informality and territorial inequality; and the need for policy portfolios rather than single instruments. Three divergences shaped feasible action: flow versus stock pollution, short versus long response horizons, and concentrated versus diffuse financing responsibility. The synthesis supports a proposed dual-track framework. Immediate measures prioritize source control, surveillance, safer work, and protection of exposed populations; longer-term measures address clean infrastructure, remediation, and the redesign of production and consumption. The profiles, horizons, and portfolio shares are interpretive planning tools, not measured effects, aggregate rankings, or optimized allocations. The contribution is the integration of established governance concepts into a Colombian comparison that connects urgent exposure reduction with equitable, sustained environmental improvement.

Student retention has become a central concern for higher education systems because it relates directly to access, equity, academic quality, and institutional efficiency. When students leave before...

This Data Descriptor presents a georeferenced dataset of electromagnetic-field exposure measurements collected in Colombia between 2 November 2023 and 30 June 2024. The records originate from the Sistema de Monitoreo de Campos of the Agencia Nacional del Espectro (ANE), which uses isotropic probes to monitor broadband radiofrequency electromagnetic fields from 100 kHz to 8 GHz and reports six-minute averages of incident power density in W/m². The comma-separated value file contains 1,286,346 timestamped records and seven source variables, including Well-Known Text (WKT) point geometry. The principal 2024 subset comprises 995,238 records from 24 monitoring stations. We provide a reproducible workflow for timestamp and coordinate parsing, structural and numeric validation, station-level aggregation, and spatial sensitivity analysis using inverse distance weighting with the mean, median, and 95th percentile. The dataset does not provide frequency-resolved measurements, calibration certificates, or measurement-uncertainty budgets; these omissions limit the use of the public file for formal regulatory-compliance assessment. The accompanying repository includes validated station-level summaries, descriptive tables, figures, and reproducible R code. The package supports environmental monitoring, geospatial analysis, methodological comparison, and reproducible reuse of electromagnetic-field exposure data.

Student retention analytics often treats withdrawal as a single outcome, although students may attribute dropout intention to personal, socioeconomic, and academic pressures simultaneously. We benchmarked nine supervised classifiers for identifying a concurrent three-domain attribution profile in a cross-sectional survey of 333 undergraduates at a Colombian public university campus. The response came from a semi-structured weight-allocation item; an audit found that literal label matching altered 21 classifications because of spelling variants and decimal notation. Nine classifiers—logistic regression, decision tree, random forest, neural network, Gaussian Naïve Bayes, k-nearest neighbors, AdaBoost, gradient boosting, and XGBoost—were fitted using six pre-specified predictors. Models were compared by repeated nested stratified cross-validation (five outer folds, three repeats), inner tuning, fold-contained preprocessing, and training-only threshold selection. The concurrent profile occurred in 256 students (76.9%). Logistic regression achieved the highest mean held-out ROC AUC (0.674, 95% CI 0.644–0.704), closely followed by random forest (0.671, 0.641–0.702); their paired difference was nonsignificant after Holm adjustment. Logistic regression had the highest F1 score (0.788), whereas random forest had the highest balanced accuracy (0.617). AdaBoost did not retain its apparent single-holdout advantage. Housing and financial aid had the largest held-out permutation importance. The predictors provided moderate discrimination of a perceptual profile, not a validated prediction of future dropout. Outcome auditing and leakage-free validation materially changed the model ranking.