top of page

Discussion

The hotspot and clusters mainly persisted in Downtown and Scarborough in most conditions, though the reasons for the formation may vary.  Downtown undertakes the role of the city’s commercial, financial and recreational center. Dense streams of migrant people and vehicles require a dense network of roads, thus results in an overall high clustered fatal traffic accidents (Figure 1). However, when we count numbers, the downtown area did not show much more large counts than other fields, the fatal crashes happened in downtown were mostly randomly distributed on different roads and intersections, presenting in maps by small but dense dots. Scarborough is featured in the industrial area and dendritic road system (Figure 2). Though the specific correlation between zoning, road structure, and traffic severity is yet not confirmed, we doubt the impact of the road system on fatality based on the similar clusters at intersections for Scarborough, Steeles, and Malton (Figure 2).  The complex road system may increase the likelihood of fatal accidents. In terms of different causations of hotspots in the two regions, our suggestions were various. The potential strategy in downtown is to reduce the road density by closing or narrowing the significant roads and controlling the traffic volume for rush hours. For Scarborough, since several intersections show a high frequency of fatal traffic accidents, the necessary monitor should be added to these dangerous intersections (Figure 2). Also, roundabouts could be an effective way to reduce the crashes at the intersection in the less urban area. 

downtown1.png

Figure 1. Toronto road systems for Downtown Toronto

scarborough.png

Figure 2. Toronto road systems for Scarborough

The interactive map below shows the complete road system in Toronto. Zoom in and check different areas in detail.

Figure 3. Interactive map of Toronto's road system

Based on the variables we determined to have significance on the proportion of fatalities amongst total overall accidents, this provided suggestions on where and what could be done to target reductions in fatalities. Between almost all groupings there was an increase in rates of speeding involvements in accidents that were fatal compared to overall accidents. This issue of speeding happens consistently at mid-blocks at both overall and fatal accidents at high levels, whereas there are increases in speeding involvement during evening rush hours and at neutral groups where all variables are equally considered. Since all these groupings saw the shift of speeding to medium high to high levels, this suggests that speeding is an important factor that separates accidents to be fatal rather than non-fatal. For evening rush hours, this can then pinpoint locations and target speeding reductions that can be more spatially applicable for commuting zones during rush hours. Areas where accidents were variables were more neutrally grouped, the increase of the inclusion of speeding involvement in fatal accidents was important as this showed how less spatially grouped areas faired between overall accidents and fatal accidents. This meant that overall speeding lead to more fatalities and these areas increased and were more distributed across the Toronto area. In this case, suggestions would be to target speeding overall as it is shown to be increase fatalities all over the study area and being less spatially predictable means less spatially targetable approaches.

Limitation and Future Direction

Our analysis only focused on the spatial distribution of where the accidents happened but did not put traffic migration into consideration. Traffic analysis, rather than other spatial analyses which can be analyzed based on neighborhood, it has a unique flowing characteristic. The fatal traffic accidents that happened in one place usually had no specific relation with its surrounding population. Crashed vehicles may come from remote areas by driving along the roads.  More data such as traffic volume was necessary to better interpret the reasons behind the fatal traffic spatial distribution,. However, due to the information opacity, we cannot access such data. 

​

The limitations of the type of data included and recorded for public access by the city of Toronto limited us in the kinds of analyses that could be done due to its categorical nature. The limited nature of categorical and binary data meant that we could not do an ordinary least squares regression or a geographically weighted regression, which we initially planned to do. 

​

Also, the research did not analyze the impacts of various road types and road density on fatal traffic accidents. Though the kernel density and road system map pointed out that the dendritic road structure may cause more fatal accidents than the interconnected road system, we do not have direct spatial analysis to prove the argument.  Future research could focus on the road structure. 

bottom of page