Analysis of urban quality of life in neighborhoods areas with emphasis on comparative methodological approaches; The case study on district 6 of Tehran

Today, with the growing population and the expansion of cities, assessing the quality of urban life is of great importance. These evaluations give organizations and city managers a good perspective for future planning and budget allocation. In the meantime, there are various methods for assessing the urban quality of life, among which the combined use of GIS-MCDA can provide an assessment of the quality of urban life with unique capabilities of GIS. 

In this study, OWA method, which is one of the multi-criteria decision-making methods with two local and global approaches, was used to evaluate the quality of urban life. OWA is a method in which multi-criteria analysis is performed based on multiple and contradictory criteria and the best solution is selected. In fact, the OWA method is a risk-based approach in MCDA. In other words, this method introduces a new concept of Boolean decision-making method and weighted linear combination that has different degrees of risk (risk-taking). The OWA operation develops a variety of strategies, from the highly pessimistic mode defined in AND to the highly optimistic mode defined as OR in Boolean logic. The risk level can be adjusted by determining the degree of ORness from 0 to 1. 

To select the evaluation criteria, first the available resources in this field were examined and then the opinions of experts were collected. Because local conditions are different and people in different places may have different criteria for the quality of life, similar criteria were combined and criteria were localized. Finally, these data were collected in six layers including: traffic, population density, access to green space, access to health services, sidewalk condition and air pollution. Then, the weights of each layer were calculated using the AHP method with an inconsistency rate of 0.08. In the second phase, raster layers with pixels of 15, 30 and 60 were obtained, each of which covers an area of 225, 900 and 3600 square meters, respectively. This change in the size of different pixels and its sizes was calculated according to the area of each neighborhood and reaching the appropriate pixel size for the study area. The reason for this variation in the pixel size of the layers produced is to study the effect of scale and achieve a scale suitable for the study area because over-shrinking the pixel size increases the cost of collecting field data while increasing processing time and data collection. Also, the large size of the pixels will cause merging and generalization and will make the ranking of places unrealistic. Therefore, the necessary care must be taken in the size of the pixels. Surveys on six neighborhoods including: Islamic Revolution, Valiasr, University of Tehran, Ghaem Magham, Behjatabad and Laleh Park and layers of air pollution, traffic, access to health services, sidewalk quality, access to green space and population density showed the use of local and global approach, Changes in pixel size and varying degrees of risk-taking can each have a direct and enormous direct effect on the results of the urban quality assessment process. With the change of each of the cases, while a significant change in the classification area in many cases occurs, the ranking of the regions also changed. In examining different approaches, using a degree of risk-taking of 0.7 and pixel sizes of 30 and 60 with a 5 * 5 filter has given better results. Therefore, due to the high sensitivity of the expressed variables and the direct effect of the weight of each layer, high care must be taken in selecting the variables and the weight of the experts. 

Comparing the two approaches, it can be seen that in the local approach, in general, "completely undesirable" areas are more scattered in the center and southeast. While in the global method, the dispersion of "completely undesirable" areas does not have a fixed cumulative pattern in the area and its effect can be seen throughout the neighborhoods in the form of small and large spots. It is evident from the results that "Fully desirable" areas in this approach are mostly seen on the outskirts of parks and boulevards. According to the results, the highest percentage of "perfectly desirable" floor area compared to the area itself is related to conditions with a degree of risk of 0.7, 3 * 3 window and a pixel size of the input layers of 15 meters.