商品多樣性指標之空間資料探勘分析：以臺北市多核心商圈為例

Spatial Data Mining of Product Variety Characteristics: Examples from the Multiple Shopping Areas in Taipei City

游舜德(Tony Shun-Te Yuo)

《都市與計劃》 49卷1期 (2022/03) Pp. 35-70

https://doi.org/10.6128/CP.202203_49(1).0002

摘要

都會區之商圈進行多樣性品質的衡量與比較，至今並無令人滿意的方法與實證，這是由於購物商圈內的商品多樣性具有豐富度、多層次，並在交互作用下的無形複雜特質所致。商品多樣性能否產生聚集經濟，在於商圈內的個別商店彼此間是否能夠產生正向的外溢效果，創造該商圈的購物氣氛效果。然而，由此難以辨識與掌握的特質，使商圈上多樣性的衡量，需要以發展出多個探勘程序來完成產生各種多樣性衡量指標。本研究提出就五種商圈多樣性的衡量特徵，除了豐富度之外，至少應該包含歧異度與均衡度、集中度以及核心主導性等主要面向，方能顯現出不同商圈各種商品多樣性的樣貌。本研究以臺北市中心為範圍，進行實地深度調查與資料建置，透過此一詳細的零售業種分布的空間資料庫，針對商圈商品多樣性的幾個根本性問題進行探討：1.客觀商圈劃設範圍的界定方法、2.商品多樣性的辨識與萃取、3.可衡量指標及其效果評量以及4.商品多樣性對零售租金的非空間與空間檢測結果，並以三維視覺化展現各種指標空間分布的情形。研究結果顯示，利用本研究所產生的不同多樣性指標，能夠提供各商圈本質上的優劣與特性，而各商圈的多樣性特徵也確實呈現高複雜的差異性。高平均租金可以來自各種多樣性變數均占絕對優勢的商圈，但實證結果也顯示低租金也可能有具有高多樣性指標；此時，核心主導多樣性的指標得點，將成為反映商圈租金水準差異的重要因素。

關鍵字

零售聚集經濟 ； 商品歧異性 ； 購物氣氛 ； 外溢效果 ； 資料探勘

並列摘要

Current research has not yet provided satisfactory methods and empirical studies to measure and compare the quality of product variety among the multiple shopping areas within an urban area, due to the complexity, intangible, and interactive nature of the subject. The increasing return from agglomeration economies reply on generating positive spillover effects from nearby stores. Hence, shopping atmosphere derives from the variety from the clustering of different retail and service stores. Nevertheless, due to the difficulty in identification the scope and magnitude, this study develops a data mining process to extract the environmental variety data. Using the GIS-based database, this study calculates and defines different variety indices, including richness, diversity, concentration and the dominant variety characteristics. This investigation adopts central Taipei city as the survey area to tackle several issues: 1) objectively identifying the shopping areas; 2) identifying and extracting the fundamental variety distributions; 3) measuring and generating various indexes representing different diversity patterns, and 4) examining the influential significance with spatial performance data, i.e. retail rents. Finally, the power of comparison among major shopping areas is displayed and enhanced using 3 dimensional visibility methods. Analytical results show that the examined variety indices can assist the understanding of spatial and non-spatial features from different perspectives. Although the dominant variety features are generally found in shopping areas with the highest average rent, some shopping areas with lower rent also have high diversity and market concentration. In this case, the core variety component that can generate increasing returns could determine retail rent rather than other variety patterns.

並列關鍵字

Retail agglomeration economies ； Product diversity ； Shopping atmosphere ； Spillover effect ； Data mining

參考文獻

Abdel‐Rahman, H. and Fujita, M. (1990). Product variety, Marshallian externalities, and city sizes, Journal of Regional Science, 30(2): 165-183. DOI: 10.1111/j.1467-9787.1990.tb00091.x.

Acar, W. and Sankaran, K. (1999). The myth of the unique decomposability: Specializing the Herfindahl and entropy measures?, Strategic Management Journal, 20(10): 969-975. DOI: 10.1002/(SICI)1097-0266(199910)20:10<969::AID-SMJ57>3.0.CO;2-0.

Allred, A. T. and Chakraborty, G. (2004). Can a marketing product classification system predict Dot. com retail success and failure?, Journal of Internet Commerce, 3(1): 31-46. DOI: 10.1300/J179v03n01_03.

Arentze, T. A., Oppewal, H., and Timmermans, H. J. P. (2005). A multipurpose shopping trip model to assess retail agglomeration effects, Journal of Marketing Research, 42(1): 109-115. DOI: 10.1509/jmkr.42.1.109.56884.

Brakman, S. and Van Marrewijk, C. (2009). Introduction: Heterogeneity at different spatial scales, Journal of Regional Science, 49(4): 607-615. DOI: 10.1111/j.1467-9787.2009.00626.x.