Skewed geography plotting census shapefiles in R

Question

I've pulled US census data by block and merged the block shapefile on it manually with the sf and rgdal packages. When I plot the resulting sf object it's showing a skewed map despite the crs remaining the same from the read-in of the shapefile through the merge.

The shapefiles can be found here: https://www2.census.gov/geo/tiger/TIGER_RD18/LAYER/TABBLOCK20/, the one for Illinois that I'm using is tl_rd22_17_tabblock20.zip.

The census data can be found here: https://data.census.gov/table?q=population&g=0500000US17031$1000000&tid=DECENNIALPL2020.P1

Here's what should be a MWE:

library(ggplot2)
library(sf)
library(rgdal)
library(data.table)

dt_census <- fread(path_to_census_data)

blck_shp <- readOGR(dsn = block_shapefile_filepath, layer = "tl_rd22_17tabblock20")

blck_sf <- st_as_sf(blck_shp)

dt <- merge(blck_sf,dt_census, by.x = "GEOID20", by.y = "GEOID")

mini_dt <- dt[,c("POP20","geometry")]
mini_dt$POP20 <- as.numeric(mini_dt$POP20)

plot(mini_dt)

but this produces a skewed plot where some blocks on the northeast are clearly outside of city lines and into Lake Michigan.