top of page

Conclusion 

​

In the exploratory regression analysis, the census variables chosen had a low adjusted R-squared value. After running an OLS and then GWR, we were able to visually assess the model’s strength and the relationship between each explanatory variable and the dependent variable. The GWR model explained a decent amount of the density of crime incidents however the relationships between the dependent variable and the explanatory variables did not entirely align with those I found through literature on socioeconomic statistical correlation of criminal behaviour.

​

The results of the study show that our model was unable to explain the all crime events with our census variables. It seems that in Vancouver, the influence of socio-economic and racial factors such as education, visible minority, and immigrants are not as evident as we expected. Socio-economic variables were limited to what the census was able to provide and given that the census information was obtained in 2011, the data has most likely drastically changed since then. Although, variables taken from the census did somewhat correlate with socioeconomic status (income, education, occupation), they were not entirely representative of the statuses. For example, housing cost was used instead of income which can be argued that both are interrelated, however, obtaining the actual income data may have been more beneficial. 

​

Altogether, the results showed that socioeconomic factors are definitely a contributor to the frequency of crimes with different weights of correlation for specific crimes. In particular, the areas of a higher adult population, lower population of postsecondary degree holders, higher population of unemployment, lower average house value are more susceptible to crime. Establishing these correlations would help with creating solutions to aid socioeconomic problems, as well as decrease crime associated to those socioeconomic factors in the meanwhile. If a more explanatory model was able to be created with high adjusted R-squared values, it may have also been possible to conduct a prediction of crime in the future given the propoer associative socioeconomic variables. Hence, for future analysis, a predictive model of crime could even be derived and carried out to ensure proper social infrastructure is prepared to deal with increased or decreased crime in particular areas. 

​

​

​

​

​

bottom of page