Hello everyone. This blog post is again guided by a prompt for our lab group. We were asked to bring something to the group that we are struggling with, so that we can work through it as a group and hopefully find some solutions. I thought a lot about what to actually use this discussion for, and I have decided to discuss a problem I have been thinking about for my statistical analysis. I'm attempting to do an analysis that requires me to ask a dataset of craters how many other craters are nearby. The problem that I am having, is that craters are not all the same size and may have overprinted some of their closest friends when they impacted. So. I have three possible options that I have been considering, and I wanted to ask the group if one or multiple of these solutions seems to best mitigate the problem, or even better, if someone has a different solution that would eliminate this problem. Here are the options:
Start with the center point of each crater. Create a buffer, a circle, with a set radius from that point outward. Ask the craters how many other crater center points are inside their circle.
Start with the rim of each crater. Create a buffer, a circle, with a set radius from the rim outward. Ask each crater how many other crater center points are inside their new circle.
Start with the center point of each crater. Create a buffer that has a radius dependent on the diameter of that crater. Ask each crater how many other crater center points are inside their new circle.
Is there another way to determine if craters are in a crater dense area? I have considered that determining how much empty space is around as opposed to how many craters are around might be one solution, but as of yet I have not tried to implement that. It again poses the problem of neighborhoods. For a given crater, how far out do I ask it to look and calculate the empty space? Hmm. Food for thought. Is it arbitrary?
Anyway, this is my question for my lab group. Another post will be coming in around two weeks. Cheers!