Nonetheless, an improvement in the regression fitting is not obtained, since the determination coefficients remain almost the same as before and are even slightly worse (R2 = 0.0312 and 0.0363). Also, these zero data prove valuable to the analysis as they allow to establish the difference between cells showing no vulnerability and the others. It would not make much sense in the scope of this study to establish a comparison and perform regression fitting only on data, which corresponds to the more deprived people.
Conclusion
The purpose of this paper was to look into the relationship between vulnerability and closeness to service stations and their related polluted areas. This consisted in an interesting exercise, since the correlation between both has not been studied in as much depth as is the case for pollution emanating directly from road networks. The results appear to confirm the initial hypothesis that more deprived people find themselves in areas closer to gas stations and their pollution. This could come as another aggravating factor for their health, as this category of people are usually already the most affected by transportation pollution. To further enlarge this study it would have been interesting to develop other models that would better approximate the data and show clearer results. Also studying larger data, for example over the entire city of Geneva, would have proved valuable in order to confirm the results of this small pool of experience.
In conclusion, the results obtained here show the subject is worth exploring. This work would be well suited as a complement to a study on road networks pollution levels and their correlation with deprivation to better encompass the entirety of the issue.