Data format for generalised linear mixed models

Question

I am very new to GLMMs and have never done one before.

Essentially, I want to understand the influence of certain predictors on the presence or absence of two mammals.I will do this in R.

Ten out of 13 of my predictor variables are in the form of a raster and they will be resampled to a 10 x 10 km grid. The other 3 are statistical data that will be manually attached to the attribute table of district (admin boundaries). Here, I will first select the districts that occur in the ranges of these two species and then manually enter the statistical information to their attribute tables.

My questions are:

1. Do I need to somehow match the district data to the 10 x 10 km grid or does the analysis occur in two levels (one where it just takes the grids and another where it uses the admin boundaries, meaning I simply retain the data at the district level)?
2. Second, for the first level, I will change the resolution of the mammal ranges to match the grids (10 x 10 km) and code each pixel as 1 or 0. I want to understand if I will use the same range map of the mammals at the 10 x 10km resolution for analysing the districts too.

I have tried to understand this via Google, however, I cannot seem to understand the levels.