The present research is dedicated to investigating the explanatory power of relative location variables in assessing and forecasting market values. Here, relative location refers to the spatial position (geographical context) of a building or property in relation to a given Point Of Interest (POI). Specifically, a methodological approach is proposed for identifying the most suitable quantification modality based on statistical performance and consistency with the market mechanisms of the specific reference context. For a case study in Northern Italy, we collected data on 615 residential properties and 2673 POIs, including cultural facilities, school and education institutions, commercial services, sports, entertainment, and leisure facilities, health and care services, public transport systems, urban parks, and green areas. The relative location between the collected properties and the POIs is assessed using an automated calculation procedure developed in the Python programming language, in conjunction with Geographic Information Software (GIS). This automatism allows the assessment of relative location in terms of different Units Of Measure (UOM), such as straight-line distance, travel time by car, travel time on foot, travel time by public transport, and the number of POIs in a 400 m/1 km ring buffer. Since 615 residential buildings and 2673 POIs were analysed, with their relative locations measured using six different UOMs, a database of 9'865'215 data was produced. Furthermore, for each category of POI, a feature importance analysis guides the selection of the best UOM, i.e., the most statistically significant one. Considering the chosen UOM, an optimised econometric technique is finally implemented to analyse the functional relationships between the market values of residential properties and the set of identified relative location variables.

中文翻译：

---

在市场价值评估中衡量相对位置变量的“最佳”方法是什么？应用于意大利案例研究的计量经济学方法


本研究致力于调查相对位置变量在评估和预测市场价值中的解释力。在这里，相对位置是指建筑物或财产相对于给定兴趣点 （POI） 的空间位置（地理环境）。具体来说，提出了一种方法，用于根据统计表现和与特定参考背景的市场机制的一致性来确定最合适的量化模式。对于意大利北部的案例研究，我们收集了 615 个住宅物业和 2673 个 POI 的数据，包括文化设施、学校和教育机构、商业服务、体育、娱乐和休闲设施、健康和护理服务、公共交通系统、城市公园和绿地。使用以 Python 编程语言和地理信息软件 （GIS） 开发的自动计算程序来评估所收集的属性与 POI 之间的相对位置。这种自动化允许根据不同的测量单位 （UOM） 评估相对位置，例如直线距离、汽车旅行时间、步行旅行时间、公共交通工具旅行时间以及 400 m/1 km 环形缓冲区中的 POI 数量。由于分析了 615 栋住宅楼和 2673 个 POI，并使用 6 种不同的 UOM 测量了它们的相对位置，因此生成了一个包含 9'865'215 数据的数据库。此外，对于每类 POI，特征重要性分析指导选择最佳 UOM，即最具统计意义的 UOM。 考虑到所选的 UOM，最终实施了一种优化的计量经济学技术，以分析住宅物业的市场价值与已确定的相对位置变量集之间的功能关系。