re-construct 70s and 80s tract ids (#15)

* re-construct 70s and 80s tract ids

* up version

* pad all tract ids to 6 char

* details on tract id construction

00_make_block_group_shp.R

@@ -24,7 +24,9 @@ saveRDS(blk_grps_sf_2000, "block_groups_2000_5072.rds")
 # 1990
 blk_grps_sf_1990 <- st_read("/Users/RASV5G/Downloads/nhgis0016_shape/nhgis0016_shapefile_tl2000_us_blck_grp_1990/US_blck_grp_1990.shp")
 
-blk_grps_sf_1990 <- st_transform(blk_grps_sf_1990, crs=5072) %>%
+blk_grps_sf_1990 <- st_transform(blk_grps_sf_1990, crs=5072)
+
+blk_grps_sf_1990 <- blk_grps_sf_1990 %>%
   dplyr::select(fips_block_group_id_1990 = STFID,
                 geometry) %>%
   mutate(fips_block_group_id_1990 = as.character(fips_block_group_id_1990))
@@ -34,8 +36,15 @@ saveRDS(blk_grps_sf_1990, "block_groups_1990_5072.rds")
 # 1980
 tracts_sf_1980 <- st_read('/Users/RASV5G/Downloads/nhgis0018_shape/nhgis0018_shapefile_tl2000_us_tract_1980/US_tract_1980.shp')
 
-tracts_sf_1980 <- st_transform(tracts_sf_1980, crs=5072) %>%
-  dplyr::select(fips_tract_id_1980 = GISJOIN2,
+tracts_sf_1980 <- st_transform(tracts_sf_1980, crs=5072)
+
+tracts_sf_1980 <- tracts_sf_1980 %>%
+  dplyr::mutate(state_fips = stringr::str_sub(NHGISST, 1, 2),
+                county_fips = stringr::str_sub(NHGISCTY, 1, 3),
+                tract_fips = stringr::str_sub(GISJOIN, 9, -1),
+                tract_fips = stringr::str_pad(tract_fips, 6, pad = "0"),
+                fips_tract_id_1980 = glue::glue('{state_fips}{county_fips}{tract_fips}')) %>%
+  dplyr::select(fips_tract_id_1980,
                 geometry) %>%
   mutate(fips_tract_id_1980 = as.character(fips_tract_id_1980))
 
@@ -44,8 +53,15 @@ saveRDS(tracts_sf_1980, "tracts_1980_5072.rds")
 # 1970
 tracts_sf_1970 <- st_read('/Users/RASV5G/Downloads/nhgis0018_shape/nhgis0018_shapefile_tl2000_us_tract_1970/US_tract_1970.shp')
 
-tracts_sf_1970 <- st_transform(tracts_sf_1970, crs=5072) %>%
-  dplyr::select(fips_tract_id_1970 = GISJOIN2,
+tracts_sf_1970 <- st_transform(tracts_sf_1970, crs=5072)
+
+tracts_sf_1970 <- tracts_sf_1970 %>%
+  dplyr::mutate(state_fips = stringr::str_sub(NHGISST, 1, 2),
+                county_fips = stringr::str_sub(NHGISCTY, 1, 3),
+                tract_fips = stringr::str_sub(GISJOIN, 9, -1),
+                tract_fips = stringr::str_pad(tract_fips, 6, pad = "0"),
+                fips_tract_id_1970 = glue::glue('{state_fips}{county_fips}{tract_fips}')) %>%
+  dplyr::select(fips_tract_id_1970,
                 geometry) %>%
   mutate(fips_tract_id_1970 = as.character(fips_tract_id_1970))
 

README.md

@@ -7,7 +7,7 @@
 ## DeGAUSS example call
 
 ```sh
-docker run --rm -v $PWD:/tmp ghcr.io/degauss-org/census_block_group:0.4.1 my_address_file_geocoded.csv 2010
+docker run --rm -v $PWD:/tmp ghcr.io/degauss-org/census_block_group:0.4.2 my_address_file_geocoded.csv 2010
 ```
 
 * The first argument (`my_address_file_geocoded.csv`) is the name of your geocoded csv file.
@@ -45,6 +45,8 @@ docker run --rm -v $PWD:/tmp ghcr.io/degauss-org/census_block_group:0.4.1 my_add
     | County | State + County | 2+3=5 | Hamilton County | 39061 |
     | Census Tract | State + County + Tract | 2+3+6=11 | Tract 32 in Hamilton County | 39061003200 | 
     | Block Group | State + County + Tract +<br /> Block Group | 2+3+6+1=12 | Block Group 1 in Tract 32 | 390610032001 |
+
+Due to inconsistencies in the 1970 and 1980 tract identifiers, we concatenated the state FIPS (`NHGISST`), county FIPS (`NGHISCTY`), and tract FIPS (the last 4 or 6 digits of `GISJOIN2`) to construct the full `fips_tract_id`. Since the length of tract FIPS codes varied, we padded all tract FIPS to the maximum 6 digits using zeros. 
 
 ## DeGAUSS details
 

census_block_group.R

@@ -1,7 +1,7 @@
 #!/usr/local/bin/Rscript
 
 dht::greeting(geomarker_name = 'census_block_group',
-              version = '0.4.1',
+              version = '0.4.2',
               description = 'adds census block group or tract identifiers to geocoded addresses')
 
 dht::qlibrary(dplyr)