Abstract
Residential land use is one of the important components in urban land, since it is an essential element of people’s lives. Residential developments in general receive more attention by residents, as well as developers and governments. However, old dwellings cannot meet current requirements of residents and obstruct further urban development. Many old residential estates need to be redeveloped or renewed. Therefore, new residential projects are constructed by real estate developers in many cities. For obtaining more profits, developers focus on the location of residential developments which can attract more consumers to purchase residential properties. Therefore identification of the optimal location for residential development projects is an important research topic.With respect to residential development, previous studies have focused on the residential value and location choice. Some studies relevant to location choice pay more attention to the influence of housing attributes on the process of residents’ decision-making for moving in a new house. However, which location among new residential development projects is optimal remains unclear. As a result, simulation of the optimal location choice for residential development projects is still being overlooked. In this study, a Residential Preference based Location Choice Model (RPLCM) is developed to simulate the optimal location of residential development project both in given areas and a selected city. The proposed model puts forward the major attributes of consumers’ residential preferences for a new house. Moreover, Residential Preference Index (RPI) which is defined as the attractiveness of residential location can be calculated by the RPLCM. Therefore, the RPLCM is able to determine the optimal location from among different alternatives which can attract maximum residents to purchase.
According to previous studies, four types of attributes with 15 sub-variables are identified as the factors which may influence the attractiveness of different residential locations. Four attributes in the first level are accessibility attributes, neighborhood characteristics, environmental factors and socio-demographic characteristics. Reliability Interval Method (RIM) is used to identify the major variables which mostly affect the residents’ preference for location. A questionnaire survey is processed to collect the data from four groups of dwellers according to categorisation of population by age in Hong Kong. Results present that three attributes play the most role in the process of decision-making by residents to purchase a residential property. They are access to metro station, access to Central Business District (CBD) and monthly median household income.
Parameters of the three major variables used in the proposed model are calibrated based on the data of 116 private residential estates in Hong Kong. The data are from Centadata, HK Population Census, Google Map and Open Street Map in Hong Kong. In this study, Iterative Integer Method is adopted to calibrate parameters after testing the Pearson’s correlation coefficient. High value of a residential development project means high profit for developers as well as optimal location of the residential project. Thus, the minimum variance shows the trend of RPI is closest to the trend of residential value. The group of parameters with the minimum variance contains appropriate parameters in the proposed model.
The proposed model is tested from two aspects: scenarios in different districts or in the same district. Different areas in Hong Kong are categorised into three districts: Hong Kong Island, Kowloon and New Territories. The RPIs of three random locations in three districts as well as RPIs of three groups of observations in each district are calculated by the RPLCM. As a result, locations which have the highest RPIs are the optimal locations suggested to developers. Furthermore, the values of these observations are collected to validate the results of the proposed model. The trend of RPIs of every group both in different districts and same districts are same as the trend of their values except two observations in Kowloon. One project in Kowloon was constructed 10 years before another. Thus, the error can be accepted, and the proposed model can be assumed to have good accuracy.
The validated RPLCM is widely applied to both district level and city level. This study uses the RPLCM to make a visual residential preference map and evaluation table of Hong Kong. The available data of 280 tertiary planning units (TPUs) of Hong Kong are collected. The RPIs which represent the attractiveness of these TPUs are calculated by the RPLCM and divided into ten levels based on quantiles. The evaluation table with ten levels can help public to identify the attractiveness of a specified TPU by searching the RPI. Moreover, the visual residential preference map according to the RPI of each TPU shows visualised residential conditions in Hong Kong.
This research develops a comprehensive framework with both reliability interval method and residential preference based location choice model to simulate the optimal location of residential development projects. The RPI can help residents decide whether and where to move in a new residential property. More importantly, this model can help developers to evaluate the attractiveness of different scenarios and identify the optimal location of residential development projects from among available alternatives. Furthermore, the visual residential preference map and evaluation table can help the government to understand the residential conditions of different districts and be the basis for future residential land use planning.
| Date of Award | 14 Mar 2018 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Siu Ming LO (Supervisor) & Young Chul KIM (Supervisor) |