近年來,中國大陸的城市化發展迅速。隨著城市化的發展浪潮,房屋交易也變得更多。典型的例子是位於浙江省杭州市的房地產市場繁榮,因此被選為研究區域。貝殼找房是鏈家這一仲介旗下的房屋交易網站,該網站會發佈其平臺人員經手的成交案,我們通過貝殼找房收集了2019年1月至2020年9月的32,000多宗二手房屋交易記錄。這些屬性包括房價,房屋類型,房屋面積,方向,建築風格,電梯,裝修,建成年代和出售時間。基於收集的資料,旨在建立一個迴歸模型來估算房地產價格。 為此,除了上面列出的屬性外,我們還收集了從遙測圖像中提取的NDVI、NDBI和LST等環境因素。通過高德地圖API獲取捷運站、中小學、公園廣場等POI點位。并且還加入距離舊CBD、距離新CBD以及距離西湖等變量。爲了驗證COVID-19以及遙測指數對於房價的影響,我們做了探索性空間數據分析。 因爲受到COVID-19大流行的影響,杭州市2020年第一季度二手房屋成交極少,隨著疫情在杭州的結束,杭州市二手房市場在第二季度迎來大增長。2020年前三季度杭州市的NDVI較去年同期有明顯增長,而LST在房屋高成交區有略微的下降,不過杭州市整體的NDBI略有上升。 最後將遙測資料獲得的指數被放入特徵價格模型中以執行迴歸計算,結果表明NDVI、NDBI和LST對於房價是顯著的。NDVI與房價正相關,而NDBI和LST則與房價負相關。
In recent years, the urbanization of mainland China has developed rapidly and the number of property transactions also increases largely. A typical example is a boom in the real estate market in Hangzhou, Zhejiang Province, so it was selected as the research area. "Shell Search" is a real estate trading website publishing the transaction cases handled by its platform. We collected about 32,000 second-handed houses transacted from January 2019 to September 2020 through Shell Search. In which the attributes, including housing price, housing type, housing area, orientation, architectural style, elevator, decoration, built year, sale time, etc., were recorded. Based on the collected data, we aimed to establish a regression model to estimate real estate prices. In addition to the attributes listed above, we also collected NDVI, NDBI, and LST extracted from remote sensing images to represent environmental factors. POI points such as MRT stations, primary and secondary schools, parks, and squares, were extracted through the Gaode Map API. We also added variables such as distance to the old CBD, the new CBD, and West Lake, in the regression model. Due to the impact of the COVID-19 pandemic, there were very few second-handed housing transactions in Hangzhou in the first quarter of 2020. With the end of the epidemic in Hangzhou, the second-handed housing market in Hangzhou ushered in significant growth in the second quarter. In the first three quarters of 2020, Hangzhou’s NDVI increased significantly compared to the same period last year, while LST declined slightly in areas with high housing transactions, but Hangzhou’s overall NDBI increased slightly. Finally, all the variables were put into the regression model. The results showed that NDVI, NDBI and LST were significant for housing prices. NDVI was positively correlated with housing prices, while NDBI and LST were negatively correlated with housing prices.