I am mapping the crash data of Austin capital area in order to find out the pattern of traffic accidents. With the X,Y coordinates, I have turned the data into points. And I downloaded the shapefile of roads. My supervisor wants me to show the crash pattern on different classes of roads, like how many crashes occured on the highway and city streets. Since there are too many points that they cluster at the same places and overlap each other, I tried to spatially join the point crash to the road shapefile. In this way, I can get the sum of points on each segment of the roads. And then, I can display the range of sum value with different corlor and size of points. My questions are:
1. The lines in the road shapefile are too thin that no points fall exactly on the lines. So when I use spatial join - "intersect", the output data is null. Then I chose to sum the "closest" points, the result is always larger than the real amount. I guess some points may be double counted and I have no idea which points would be counted as "closest" . Does anyone encounter this problem before? I would appreciate if you could share your experience with me to solve it. I am wondering if there is a way to change the thickness of the lines so the points can intersect the lines? Is there any other way to show the crash density (Except the interpolation methods like IDW)?
2. How to make maps readable? I never think about this question when I was having GIS class at school. But now in real work, I found that due to the large amount of data and number of variables, the maps I made often look messy. For instance, I need to show the density of crash and the injury severity of these crashes at the same time. Any book or suggestion is appreciated.
Thanks for your patience.