以開放大數據資料來源進行不動產交易、人口特徵與優勢資源歧異度之多層次空間探勘：臺灣本島349鄉鎮市區之不均現況分析

Abstract

開放資料與巨量資料探勘概念的提出已有相當時間，然至今尚未有能涵蓋全臺灣不動產交易與特徵資料的整合性探勘結果，對空間與非空間資訊提供充分且完整的描述與呈現。過去的文獻顯示不動產價量的決定因素，與所屬空間的人口特徵及各種優質資源，具有一定關係；故欲詳細描繪不動產交易價量的空間分布，這應該是多層次的分析，且須運用各種不同演算程序與方法達成；本研究即提出一整合複雜資料多演算法的資料探勘方法，透過開放資料來源，針對臺灣本島349個鄉鎮市區，涵蓋超過五十萬筆資料，萃取出高、中、低指標分布區，並以視覺化的方式呈現。結果顯示，臺灣當前的不動產交易、人口與資源分布確實是不均的。

While the concept of open and big data has been around for some time, scholarship on comprehensive mining results on property transaction and related spatial and non-spatial characteristics variables have not been explored or published. Previous research suggests that the determinants of property price were closely related to demographic and the diversity of beneficial resources, such as infrastructures and environmental amenities. As a result, in order to illustrate these determinants in detail, our analysis used multiple algorithms to extract these comprehensive features. From various open data sources, this study collected over half-million data numbers on the 349 town-level districts in Taiwan main island. The clusters of high, mid- and low-transaction, population and resources districts were extracted and displayed in 2D and 3D maps. The results show that spatial inequality still exists in Taiwan.