Page 50 - JCAU-6-3

Journal of Chinese
Architecture and Urbanism Utilization of rural heritage

In terms of the spatial distribution of visual environment (positive correlation), and CVI (negative correlation). The
data, a strong spatial continuity exists between the absolute values of the γ range between 0.4 and 0.7 (ESPL:
GVI and CVI, delineating a circular-layer change 0.6; R&PVI: 0.47; CVI: −0.56), with p-values showing
(Figure 8E and F). Specifically, the GVI demonstrates an significance (p < 0.05). In addition, a high degree of
increase from the northwest to the southeast, while the collinearity is observed between certain variables, such as
CVI displays a corresponding decrease. The SVI within the RH and T (γ = −0.96) and R&PVI and GVI (γ = −0.86).
a
streets of the historic town is limited, with the street’s aspect Furthermore, Figure 9 presents the VIF values for the
ratios (D/H: ratio of street width [D] to building height nine variables, revealing that the VIF values of T , RH, and
a
along the street [H]) recorded at a range of 1.3 – 2.0. The CVI exceed 7.5. Considering the results of the Pearson
R&PVI in the southwest of the historic town is relatively correlation analysis and VIF values, the variables included
high (SP16 – SP22) (Figure 8G). A newly constructed street in the regression analysis are LS, RH, ESPL, WS, GVI, SVI,
with an antique appearance features wider roads than the and R&PVI, with LS designated as the dependent variable
historic street, accompanied by sidewalks on both sides. and the others as independent variables.
The houses lining this street closely resemble those in the Table 1 presents the results derived from the model
historic street (Figure 8H). Typically, buildings in this area equation (Equation I):
consist of 2 – 3 floors. In addition, two large squares, SP18
and SP21, serve as popular gathering places for residents of LS = 42.271 – 0.110 × RH + 0.335 × ESPL – 1.722 ×
the historic town after nightfall. WS – 0.010 × GVI – 0.072 × SVI – 0.011 × R&PVI (V)
The R value of the model stands at 0.545, indicating that
2
3.2. Relationship between environmental and stress RH, ESPL, SVI, and R&PVI collectively explain 54.5% of the
data variation in LS. The F-test conducted on the model confirms
Figure 9 illustrates the results of the Pearson correlation its significance (F = 3.787; p = 0.012 [<0.05]), indicating
analysis using LS, T , RH, ESPL, WS, GVI, SVI, R&PVI, and that at least one of the independent variables affects the LS
a
CVI as variables. Statistically significant correlations are relationship. Furthermore, a multicollinearity test reveals
evident between LS and ESPL, as well as between R&PVI that all VIF values in the model are below 5, confirming
and CVI, thereby affirming the validity of the selected the absence of multicollinearity issues. The Durbin–
indicators. Notably, significant associations are observed Watson statistic indicates no autocorrelation issues in our
between LS and ESPL (positive correlation), R&PVI model, as its value hovers around 2, indicating a lack of

Figure 9. Pearson correlation coefficient between the nine variables and variance inflation factor (VIF) values of variables. Notes: CVI: Construction
view index; ESPL: Equivalent sound pressure level; GVI: Green view index; LS: Level of stress; R&PVI: Road and pavement view index; RH: relative
humidity; SVI: Sky view index; Ta: Air temperature; WS: Wind speed. Source: Graph by the authors

Volume 6 Issue 3 (2024) 11 https://doi.org/10.36922/jcau.2481