*4.1. Index Analysis*

The index analysis was developed using geographic information system (GIS) data obtained from the Jeddah Municipality, and the data were exported to an AutoCAD file. Areas within a 200 m radius (an area of approximately 12.56 ha) for mosques in villa-type neighborhoods (103 in total) were analyzed in Jeddah (Figure 4).

The index was subdivided into private and public spaces. Private spaces comprised elements such as villas, commercial facilities, and office buildings. Public spaces comprised facilities such as a mosque, an elementary school, a street, and parks. The following six parameters were established for analyzing villas: (1) numbers of villas in 2008 and 2016; (2) total areas of villas (average and median values and minimum and maximum values and the differences between them); (3) the minimum, maximum, and median values for the distances of existing villas' lot frontage and their proportion to the lot length; (4) the number of villas facing 1, 2, 3, and 4 streets; (5) the average land value in the SAR (per square meter); and (6) the total area for commercial facilities and office buildings. Nine parameters were selected to analyze public space: (1) the total area of streets; (2) the total road length; (3) the road width (maximum and minimum values, and the difference between them, and the median value); (4) the distance from a mosque to the nearest mosque; (5) the area of land occupied by the mosque; (6) the distance to the nearest elementary school; (7) the number of parks, and the distance to the nearest park and the park lot area; (8) the total area of the vacant lot; and (9) the total area of public facilities, and public parking.

#### *4.2. The PMPRLS and Actual Condition*

A scatterplot was created to analyze and clarify the following issues: (1) the extent to which the PMPRLS is applied, and (2) the typology of villa-type neighborhoods to clarify the extent to which the PMPRLS is applied. The correlation coefficients for several factors relating to the PMPRLS were calculated by performing regression analysis for existing conditions, the results of which are shown in Figures 9–12.

**Figure 9.** Street total area and average of land lot area.

**Figure 10.** Street total area and total street length.

Figure 9 shows that only 32% of the study sites are within the standard street total area (STA), whereas 68% exceed the maximum standards. Moreover, 73% of the average of villa land lot area (AoVLA) is between 500 and 1000 m2. Moreover, the correlation between the STA and AoVLA was weak for existing conditions, as shown by the value of *R*, which was 0.00815 (where a perfect correlation has a value of 1).

Figure 10 shows that only 11.5% of the study sites were within the STA and TSL (total street length) standards, with 88.5% of the sites exceeding the maximum standard. Furthermore, the TSL in 77.7% of the study sites was found to be between 1500 and 2500 m. As shown in Table 4, there was a perfect correlation between the AoVLA and the TSL. However, 94% of the study sites exceeded the standards curve (Figure 11).

**Figure 11.** Total street length and average villa lot area.

**Figure 12.** Number of villas and distance to park.

Figure 12 shows the final PMPRLS factor, which is the correlation between the number of villas and the distance to the nearest park. According to the PMPRLS, one children's playground must be provided for every 20 villas. However, our results indicated that the number of children's playgrounds (NoCP) was not applicable in reality (Table 3); therefore, we decided to investigate the distance to the nearest park instead. Two criteria were established: parks within a 200 m distance and those within a 500 m distance according to the distance recommended by the World Health Organization (WHO). In 24% of the sites, the distance to the park was between 500 and 1000 m. It was clarified that only 22% of study sites have a distance to the nearest park of 200 m which is based on the PMPRLS; however, only 41% of study sites met the recommendation of WHO regarding the distance to a park, which is 500 m. In addition, 24% of study sites are far away from parks, having a distance between 500 and 1000 m. Finally, 13% of study sites have a distance of more than 1000 m to the nearest park.

#### *4.3. Contemporary Villa-Type Neighborhoods and the Housing Shortage*

As shown in Figure 13, the number of villas in all of the study sites increased but at different rates. This finding indicates the need to examine the extent to which contemporary neighborhoods can respond to the problem of the housing shortage, which amounts to 80,000 units (Figure 6, Shortage 2). To examine this question, we assumed that the occupation pattern for the total vacant area in each site would match the growth pattern. The calculation was performed by determining the total vacant area (m2) and the average lot area of already-built villas in each study site. The following formula was applied for each study, with the results shown in Table 5:

$$f(\text{total number of villlas}) = \frac{\text{TVA}(\text{total vacant area})}{\text{AVLA} \text{ (average of villa lot area)}}$$

where *f* denotes the total number of villas that can be added to the total vacant area in the same tendency. *TVA* denotes the total vacant area (m2), and *AVLA* denotes the average of villa lot area (m2). *f* includes a decimal point in the number of villas because *AVLA* is a projection of future development. *TVA* is defined as the total area to be built on, excluding roads, within an area of 12.56 ha, which is based on a radius of 200 m centered on each mosque.

**Figure 13.** Numbers of villas in 2008 and 2016, and the ratio of vacant lots.

**Table 5.** Number of villas that can be added to a vacant lot area in each case study in a scenario entailing the same growth pattern.



**Table 5.** *Cont.*

Total number of villas that can be added to the total vacant area in all of the sites in a scenario entailing the same growth pattern 2843 units

> CSN = the case study number, and *f* = the total number of villas that can be added to the total vacant area in a scenario entailing the same growth pattern.

> The results of the calculations shown in Table 5 clearly indicate that, within contemporary villa-type neighborhoods, the pattern of full occupation by villas is the same for each case study, with neighborhoods only meeting 3.5% of the total housing shortage (2843 units/80,000 units × 100). Given this finding, in addition to several differences between the case studies relating to aspects such as land price, average lot area, the net of the total vacant lot, and others, it was necessary to clarify the indicators (mentioned above in the index analysis) that demonstrated a strong correlation via correlation analysis. The indicators were then categorized into groups according to the results of the multiple linear regression analysis and scatter plot, as discussed in the following section.

#### **5. Elements of Neighborhoods and Their Correlations**

#### *5.1. Correlation Analysis*

To contribute to addressing the housing shortage problem, it is necessary to predict the growth possibilities of villa-type neighborhoods starting from the present. This prediction was performed by analyzing the characteristics and composition of contemporary villatype neighborhoods. In order to understand the correlation between housing development trends, and the site and surrounding facilities over an 8-year period, several factors, such as the number of development sites, site area and proportion of lots, and distance between mosques and highways in 2008 and 2016, along with land prices, a factor also considered to be important in previous studies, were adopted to perform correlation analysis.

The results obtained for the increase in the number of villas within the 8-year period revealed a negative correlation of −0.447 between the land price and the increased number of villas. This result indicates that there was a higher increase ratio for the number of villas in sites where land prices were lower. However, the correlation for lots that ranged in area between 300 and 400 m2 was stronger, with a positive correlation of 0.548. Thus, there was a higher possibility of growth for small villa lots that was strongly related to the possibility of purchasing these lots. In addition, the proportion of the lot's frontage and depth was more than 2:1 for lots with areas within a range from 300 to 400 m2, with a positive correlation of 0.468 (Table 6).


**Table 6.**

Correlation

 analysis among the several factors during an 8-year period.

Furthermore, over the 8-year study period, the number of villas was positively correlated (0.480) with the distance between the mosque and the highway. However, there was a negative correlation (−0.501) for the distance between the mosque and the highway and the average land price, which means that the distance between the highway, and the mosque had the effect of increasing the land price, with greater proximity to the highway corresponding to a higher land price. In addition, there was a positive correlation (0.481) between the average land price and the road area. However, among the sites evidencing growth within the 8-year study period, those with limited street areas and low prices showed a significant increase in the number of villas.

In addition, the results of the analysis showed a positive correlation (0.458) between the increased number of villas in 2016 and the ratio of lots ranging in area between 500 and 600 m2. However, this correlation was lower than the value obtained in 2008 (by 0.037) of 0.495. Furthermore, the analysis revealed a striking relationship between lots that faced one road and the number of villas that increased between 2008 and 2016, indicated by a positive correlation of 0.523. As noted above, between 2008 and 2016, the majority of the new villas were constructed with lots ranging between 300 and 400 m2, and a decreased correlation value was found for lots having areas between 500 and 600 m2. Therefore, it is clear that most of the contemporary lots ranging in size between 500 and 600 m2 were built before 2008.
