censusdata_merge

library(tidycensus)
library(tidyr)
library(sf) 

#slow

data <- read.csv("columbus.csv")
data <- sample_n(data, 500)

coord <- data.frame(lat = c(data$hlat), long = c(data$hlong))

coord$census_code <- apply(coord, 1, function(row) call_geolocator_latlon(row['lat'], row['long']))
coord


#fast

data <- read.csv("columbus.csv")
set.seed(12345)
data <- sample_n(data, 2500)

data <- select(data, hlong, hlat, county, tract, price_000s)

col_income <- get_acs(
  geography = "tract",
  state = "OH",
  county = "Franklin",
  variables = c(poverty = "B19013_001"),
  year = 2010,
  geometry = TRUE)

col_income$row <- as.numeric(rownames(col_income))

#boundary <- st_bbox(col_income)

#data <- subset(data, hlong >= boundary[1] & hlong <= boundary[3] & hlat >= boundary[2] & hlat <= boundary[4])

coord <- data.frame(lat = c(data$hlat), long = c(data$hlong)) %>%
  st_as_sf(coords = c("long", "lat"),
           crs = st_crs(col_income))

coord$row <- as.numeric(st_within(coord, col_income))

data <- cbind(data, coord)
data$geometry <- NULL

#col_income_2 <- st_join(data, col_income, join = st_nearest_feature, by=row)

new_data <- merge(data, col_income, by="row", all.x = TRUE)

new_data <- st_as_sf(new_data)

ggplot(col_income) +
  geom_sf() +
  geom_sf(data = coord, aes(color = is.na(row)))

ggplot(new_data) +
  geom_sf() +
  geom_sf(data = coord, aes(color = is.na(row)))