I try to give my opinion about the problem, which I do not think is a duplicate of the question pointed out by @15Step in his/her comment.
First of all, two little disclaimers:
The above link will take you to my earlier question, which got a useful and sensible reply by @whuber. His reply entailed the use of a simulation in R to build the distribution of average minimum distances between individuals and buildings under the null hypothesis that there is no association between them. The latter is achieved (in whuber's example) by keeping the location of the two elements (i.e., individuals and building) unchanged and randomly swapping the assigment of each location to either of the two elements. Then the observed average distance is set against the randomized distribution and its statistical significance is computed.
Another approach would be using the G-cross function (out of the
spatstat package), as described here (again by whuber): https://stats.stackexchange.com/questions/38013/interpretation-of-spatial-gcross-plots.
Yet another index, developed in archaeology, to test for spatial association between two point features is the Hodder-Okell's A index. The A index is about equal to 1 when the two patterns are randomly mingled; it is smaller than 1 when the two patterns are segregrated; it is larger than 1 when the features of the two point patterns tend to occur together. Statistical significance is assessed via a permutation approach, not dissimilar from the one described by whuber in his reply to my earlier question. The A-index is implemented in the
Aindex() function out of a package that can be installed into R from GitHub (https://github.com/gianmarcoalberti/GmAMisc).
In summary, while I do know that I have addressed just part of your question, I think that one of the described solutions would allow you to achieve what you're after. Importing your shapefiles into R and putting one of the above solution to work would not take you too long.