Dinamika Spasial-Temporal Tenaga Kerja Transportasi dan Pergudangan Indonesia dengan Geographically and Temporally Weighted Regression

Abstract

Perkembangan gig economy berbasis platform digital telah mendorong perubahan pada sektor transportasi dan pergudangan di Indonesia, terutama selama dan setelah pandemi COVID-19. Perubahan pola mobilitas, digitalisasi, dan aktivitas logistik berpotensi menimbulkan heterogenitas spasial dan temporal antarwilayah. Oleh karena itu, digunakan metode Geographically and Temporally Weighted Regression (GTWR) yang mampu mengakomodasi perbedaan pengaruh faktor-faktor penentu menurut lokasi dan waktu. Penelitian ini bertujuan menganalisis pengaruh Indeks Pembangunan Teknologi Informasi dan Komunikasi (IP-TIK), Produk Domestik Regional Bruto (PDRB), Upah Minimum Provinsi (UMP), kepadatan penduduk, dan panjang jalan terhadap persentase pekerja sektor transportasi dan pergudangan di 34 provinsi Indonesia selama 2018–2023. Sebelum pemodelan, data ditransformasi untuk memperbaiki karakteristik data. Hasil menunjukkan adanya indikasi heterogenitas spasial dan temporal yang didukung oleh uji Breusch–Pagan serta eksplorasi boxplot. Kernel adaptive bisquare dengan bandwidth optimum 22 menghasilkan model terbaik dengan nilai cross validation sebesar 26,30 dan Pseudo R square sebesar 94,29%. Model GTWR juga memberikan performa lebih baik dibandingkan regresi linier global. Temuan ini menunjukkan pentingnya kebijakan transportasi dan logistik yang mempertimbangkan karakteristik wilayah serta dinamika waktu.

The development of the platform-based gig economy has driven changes in Indonesia's transportation and warehousing sector, particularly during and after the COVID-19 pandemic. Changes in mobility patterns, digitalization, and logistics activities may create spatial and temporal heterogeneity across regions. Therefore, Geographically and Temporally Weighted Regression (GTWR) was employed to accommodate differences in the effects of determining factors across locations and time. This study aims to analyze the effects of the Information and Communication Technology Development Index (IP-TIK), Gross Regional Domestic Product (PDRB), Provincial Minimum Wage (UMP), population density, and road length on the percentage of workers in the transportation and warehousing sector across 34 Indonesian provinces during 2018–2023. Before modeling, the data were transformed to improve their characteristics. The results indicate spatial and temporal heterogeneity, supported by the Breusch–Pagan test and boxplot exploration. The adaptive bisquare kernel with an optimum bandwidth of 22 produced the best model, with a cross-validation value of 26.30 and a Pseudo R square of 94.29%. The GTWR model also outperformed global linear regression. These findings highlight the importance of transportation and logistics policies that consider regional characteristics and temporal dynamics.