Implementing a mass valuation application on interoperable land valuation data model designed as an extension of the national GDI

Abstract

The main purpose of this study is to propose an interoperable land valuation data model for residential properties as an extension of the national geographic data infrastructure (GDI) and to make mass valuation process applicable with the use of machine learning approach. As an example, random forest (RF) ensemble algorithm was implemented in Pendik district of Istanbul to evaluate the prediction performance by using thematic datasets compatible with the data model. This study provides a methodology for various urban applications and robustness of the algorithm increases the prediction of the real estate values with the use of qualified datasets.

KEYWORDS:

• Geographic data modelling
• Data interoperability
• Mass valuation
• Machine learning
• Random forest
• Urban land management

Supplemental data

Supplemental data for this article can be accessed at https://doi.org/10.1080/00396265.2020.1771967

Additional information

Funding

This work was supported by TUBITAK [Grant Number 116Y204 numbered research project].

Notes on contributors

Arif Cagdas Aydinoglu

Prof. Arif Cagdas Aydinoglu (PhD), as geomatics engineer, has expertise on GIS, geographic data management, and Urban GIS applications (http://arifcagdas.com/en/).

Rabia Bovkir

Rabia Bovkir (MSc), as geomatics engineer and PhD student, continues her research on land valuation and big data management within smart city concept.

Ismail Colkesen

Assoc. Prof. Ismail Colkesen (PhD), as geomatics engineer, has expertise on Remote Sensing and machine learning techniques.