Replication Data and Code for “Demographic Pressure, Labour Structure and Unemployment in Africa and Europe: A Comparative Machine Learning Study”

This dataset supports the study titled “Demographic Pressure, Labour Structure and Unemployment in Africa and Europe: A Comparative Machine Learning Study.” The study examines how demographic pressure, labour structure, macroeconomic performance, education, and urbanization predict unemployment across African and European economies. The repository contains the processed analytical dataset and Python code used for the empirical analysis. The data cover Africa and Europe from 1985 to 2024, producing 4,120 country-year observations. The target variable is total unemployment as a percentage of the total labour force, based on the modelled International Labour Organization estimate. The modelling predictors include age dependency ratio, old-age dependency ratio, young-age dependency ratio, population ages 0 to 14, population ages 15 to 64, population ages 65 and above, employment in agriculture, employment in industry, employment in services, GDP growth, tertiary school enrolment, and urban population share. Continent, subregion, income level, and lending type are retained as categorical features in the modelling workflow. The accompanying Python notebook reproduces the full analysis. It loads the topic-specific Excel dataset, generates descriptive statistics, checks missingness, produces a correlation heatmap, and plots unemployment trends by continent or subregion. It then estimates Random Forest, Extra Trees, and Gradient Boosting regression models. Model performance is assessed using mean absolute error, root mean squared error, and R-squared. The notebook also generates permutation feature importance, prediction diagnostics, principal component analysis, MiniBatch K-means clustering, two-dimensional PCA cluster maps, three-dimensional PCA visualization, and cluster profile tables. The files are shared to support transparency, replication, and further research on unemployment, demographic pressure, labour-market structure, education, urbanization, and machine learning applications in comparative development research.本数据集支撑题为《非洲与欧洲的人口压力、劳动力结构与失业：一项比较机器学习研究》的学术研究。该研究探讨了人口压力、劳动力结构、宏观经济表现、教育水平与城市化水平如何预测非洲及欧洲各经济体的失业状况。 本仓库包含用于实证分析的经处理分析数据集与Python代码。数据覆盖1985年至2024年的非洲与欧洲地区，共计4120条国家-年度观测样本。目标变量为以国际劳工组织（International Labour Organization）模型估算值为基准的、占劳动力总人口百分比的总失业率。 建模所用预测变量包括：总抚养比、老年抚养比、少年抚养比、0至14岁人口数、15至64岁人口数、65岁及以上人口数、农业就业人数、工业就业人数、服务业就业人数、GDP增长率、高等教育入学率以及城镇人口占比。建模流程中保留了大洲、次区域、收入水平与借贷类型作为分类特征。 配套的Python笔记本可复现完整分析流程：加载主题专属的Excel数据集，生成描述性统计量，检测缺失值，绘制相关热力图，并按大洲或次区域绘制失业趋势图。随后将拟合随机森林（Random Forest）、极端树（Extra Trees）与梯度提升回归模型。模型性能通过平均绝对误差、均方根误差与R平方进行评估。该笔记本还可生成置换特征重要性、预测诊断结果、主成分分析、小批量K均值（MiniBatch K-means）聚类、二维PCA聚类映射图、三维PCA可视化结果以及聚类特征表。 本数据集及相关文件公开共享，旨在为失业、人口压力、劳动力市场结构、教育、城市化领域的研究，以及比较发展研究中的机器学习应用提供透明化、可复现性支持，并推动相关后续研究。