نبذة مختصرة : 학위논문 (석사) -- 서울대학교 대학원 : 행정대학원 행정학과(정책학전공), 2021. 2. 고길곤. ; This study aims to analyze factors affecting open data initiatives in Korean local governments which play a critical role in providing open government data. Specifically, a set of factors were analyzed in terms of interactions among local governments, the publics demand for open government data, local government capability and the socio-economic environment. To examine spatial interactions, a spatial econometric model known as a spatial lag model was used instead of an ordinary least squares (OLS) regression model. A cross-sectional study design was used to measure the level of data openness across the 223 local governments in Korea. Data were collected in September 2020. The level of data openness was measured by the number of variables in datasets. With the help of open APIs and Python, the number of variables were extracted from all files and APIs provided by the local governments on the Korean Open Data Portal and Seoul Open Data Plaza. Collected data were combined in a geographic information system named QGIS and statistical analysis was performed using GeoDa, a spatial analysis tool and spdep, a spatial analysis package in R. This paper presents some key findings. First, the spatial lag model was better than the alternative OLS regression model. The OLS model satisfied the assumption of normality but it violated the assumptions of variance homogeneity and independence because of spatial autocorrelation. The spatial lag model satisfied all assumptions. And some variables were not significant in the spatial lag model even though they were significant in the OLS model. Second, horizontal interactions among local governments had a statistically significant positive effect on the level of data openness. When the average level of data openness in surrounding areas increases by 1%, the level in a particular area increases by 0.4438%. Third, vertical interactions between a local government and a metropolitan government did not have a significant effect on ...
No Comments.