library(sf)
library(FNN)
library(dplyr)


setwd("")


# --- Inputs (files should be in your working directory) ---

footprints <- st_read("Footprints_county_sample.shp", quiet = TRUE)
schools    <- st_read("EDGE_GEOCODE_PUBLICSCH_2122.shp", quiet = TRUE)

# Align CRS: transform schools to footprints' CRS
schools <- st_transform(schools, st_crs(footprints))

# If the CRS is geographic (degrees), reproject both to a projected CRS (meters) for correct distances
# Here we use EPSG:5070 (NAD83 / Conus Albers) which is good for CONUS-wide work.
if (sf::st_is_longlat(footprints)) {
  target_crs <- 5070
  footprints <- st_transform(footprints, target_crs)
  schools    <- st_transform(schools, target_crs)
}

# Ensure an 'index' exists
if (!"index" %in% names(footprints)) {
  footprints$index <- seq_len(nrow(footprints))
}

# Centroid coordinates (matrix of x,y)
foot_xy    <- st_coordinates(st_centroid(footprints))
school_xy  <- st_coordinates(st_centroid(schools))

# 1-NN: for each footprint, distance to the nearest school (same units as CRS; meters if projected)
nn <- FNN::get.knnx(data = school_xy, query = foot_xy, k = 1)

# Assemble output table: index + distance
Foot_school_distances <- tibble::tibble(
  index = footprints$index,
  Foot_school_distances = as.numeric(nn$nn.dist)
)

# Save result
save(Foot_school_distances, file = "Foot_school_distances.RData")