So my project is about finding ''Road accident Hot Spots in the city and how safe are School zones. I have road accident points from 2000 to 2012 with the following fields.
crash road, crash id, crash dist, crash date, crash time, crash day of the week, speed limit, light, road wetness and case of the crash which is in coded with alphabets and numeric.
I have decided to do hotspot analysis using Optimized Hotspot analysis Tool for individual year for recent 5 years, and then for school opening and closing times. Since I am cutting down the data to certain time frame I have few incidents to work with and the hotspots end up in city centre which is not unusual. is there any way I can relate school zones to accidents? not visually though.
I presume you have output of hotspot as a point layer and school zone as a polygon. You can use spatial join tool in Analysis toolbox inside overlay toolset. It will give you count of accidents in each school zone.
You could use the Generate Near Table tool to create a table showing distance relationships between the schools and the accidents. That table could be summarized for each school showing a numerical relationship.
I am going to tag the Transportation space to see if anyone there has some additional input.
Does your schools zone reflect the temporal nature of the incident? Do your crash dates and times coincide with school hours and days or events? Also do you detect any chain reaction accidents ie... where accidents happen within a buffer zone and time of each other? I know there has been some work in NY: try streetsblog.org and Ben Wellington
I have extracted the incidents for school term during 8 to 9 and 2:30 to 3:30, used 150m buffers around schools to find number of incidents within. I found 8 schools with 1 or 2 incidents for each year but only few schools are close to accident hotspots.
One thing I don't see is the location of the accident itself. Or is that implied with the 'crash road' attribute? If you don't know where the accidents are, how can you create a hotspot map?
What type of road accidents are you looking at? Are they pedestrian-vehicle or vehicle-vehicle accidents, and are you only looking at accidents involving children? I ask because these different types of accidents have different causal factors. There are strong links between child pedestrian accidents and school locations but previous studies (e.g. The impact of the ‘school run’ on road traffic accidents: A spatio-temporal analysis) into the location of vehicle-vehicle and their proximity to schools have not found a significant relationship.
Perhaps if you are able to filter your crash data to only incidents involving children or child pedestrians you may see a trend with school locations. If this is possible another thing you may want to look at is school enrolment numbers, if you can find this data, and then use this to see if schools with more pupils have greater accident rates in close proximity, as exposure to risk is one of the main influences on the possibility of an accident occurring (i.e. higher number of children on the road = higher chances of an accident occurring). This information could then be used to help interpret the number of accidents in each school zone.