Multidimensional and Temporal Segmentation of Productive Scale in Colombian Transitory Crop Production: Department–Crop Panel Dataset (2007–2024)

This dataset contains the complete multidimensional and temporal segmentation of productive scale for Colombian transitory crop production at the department–crop level for the period 2007–2024. The data are derived from the Evaluaciones Agropecuarias Municipales (EVA), published by the Colombian Ministry of Agriculture and Rural Development (MADR), and include semi-annual observations aggregated at the department–crop level. The dataset extends a previously published segmentation dataset by incorporating the full segmentation structure across multiple analytical dimensions, including: - Productive scale clustering (large, medium, small) based on K-means.- Temporal dynamics and structural behavior of each series.- Engineered features used in the modeling process (lags, rolling statistics, and ratios).- Indicators of production intensity and variability.- Temporal and spatial concentration metrics (e.g., Herfindahl–Hirschman Index).- Classification of each department–crop pair according to its structural and temporal profile. Each record corresponds to a department–crop pair and includes both original variables (planted area, harvested area, production, yield) and derived features used for segmentation and modeling. This dataset is designed to support:- Reproducibility of clustering-based segmentation analyses.- Comparative studies of productive scale and structural heterogeneity.- Time series modeling and forecasting applications.- Policy analysis and decision-making in agricultural systems. Compared to the base segmentation dataset, this version provides a complete representation of the segmentation pipeline, including feature engineering and temporal characterization, enabling deeper analytical exploration. A supplementary file may include cluster labels and descriptive statistics for each segment.本数据集涵盖了2007年至2024年间哥伦比亚短期作物（transitory crop）生产在省（department）-作物维度下完整的多维度与时序化生产规模划分成果。 数据源自哥伦比亚农业与农村发展部（Ministry of Agriculture and Rural Development, MADR）发布的《市级农业评估报告》（Evaluaciones Agropecuarias Municipales, EVA），包含以省-作物为聚合单位的半年度观测数据。 本数据集在既往发布的划分数据集基础上进行了拓展，纳入了多分析维度下的完整划分结构，具体包括： - 基于K均值聚类（K-means）的生产规模分级（大型、中型、小型）； - 各序列的时序动态与结构特征； - 建模流程中使用的工程化特征（滞后项、滚动统计量与比率指标）； - 生产强度与变异性指标； - 时空集中度指标（例如赫芬达尔-赫希曼指数（Herfindahl–Hirschman Index））； - 依据结构与时序特征对每个省-作物组合进行分类。 每条记录对应一个省-作物组合，包含原始变量（种植面积、收获面积、总产量、单产）以及用于划分与建模的衍生特征。 本数据集旨在支撑以下研究与应用场景： - 基于聚类的划分分析的可复现性研究； - 生产规模与结构异质性的比较研究； - 时序建模与预测应用； - 农业系统中的政策分析与决策制定。 相较于基础划分数据集，本版本完整呈现了划分全流程，包括特征工程与时序特征刻画，可支持更深入的分析探索。 附属补充文件可包含各聚类簇的标签与各聚类簇的描述性统计量。