Provincial-Level Socioeconomic and Electoral Data from Türkiye's 2023 Presidential Election

This dataset contains provincial-level data from Türkiye’s 2023 presidential runoff election (May 28, 2023) integrated with a comprehensive set of socioeconomic indicators.The dataset includes 81 provincial observations and 29 explanatory variables covering economic, demographic, social, and labor-market dimensions.Official electoral data were obtained from the Supreme Election Council of Türkiye (YSK), while socioeconomic indicators were collected from the Turkish Statistical Institute (TURKSTAT), the Ministry of Treasury and Finance, the Ministry of Trade, and the Social Security Institution (SGK).Variables include GDP per capita, unemployment rate, industrial and service sector shares in GDP, education levels, migration rates, car ownership, and others.The dataset was used in the study “Socioeconomic Determinants of Voting Behavior in Türkiye’s 2023 Presidential Election: A Machine Learning Analysis” to model the relationship between socioeconomic structure and provincial voting outcomes using machine learning algorithms (Random Forest, SVM, KNN, etc.). All variables were standardized using z-score normalization, and feature selection was performed with Recursive Feature Elimination and Cross-Validation (RFECV).The data are fully anonymized, reproducible, and suitable for comparative or replication studies in political science, public policy, and behavioral economics.本数据集收录土耳其2023年总统决选（2023年5月28日）的省级层面数据，并整合了一套全面的社会经济指标。本数据集包含81个省级观测样本与29个解释变量，覆盖经济、人口、社会及劳动力市场维度。官方选举数据取自土耳其最高选举委员会（Supreme Election Council of Türkiye，YSK），社会经济指标则采集自土耳其统计研究所（Turkish Statistical Institute，TURKSTAT）、财政与财政部、贸易部以及社会保险机构（Social Security Institution，SGK）。变量涵盖人均国内生产总值、失业率、GDP中工业与服务业占比、受教育水平、移民率、汽车保有量等。本数据集被应用于题为《土耳其2023年总统选举投票行为的社会经济决定因素：机器学习分析》的研究，该研究通过机器学习算法（随机森林（Random Forest）、支持向量机（SVM）、K近邻（KNN）等）构建社会经济结构与省级投票结果之间的关联模型。 所有变量均通过z分数标准化（z-score normalization）完成标准化处理，并采用递归特征消除与交叉验证（Recursive Feature Elimination and Cross-Validation，RFECV）开展特征选择工作。本数据集已完全匿名化，具备可复现性，适用于政治学、公共政策及行为经济学领域的比较研究或重复研究。