zipcodeR is an all-in-one toolkit of functions and data for working with ZIP codes in R.
This document will introduce the tools provided by zipcodeR for improving your workflow when working with ZIP code-level data. The goal of these examples is to help you quickly get up and running with zipcodeR using real-world examples.
First thing's first: zipcodeR's data & basic search functions are a core component of the package. We'll cover these before showing you how you can implement this package with a real-world example.
The package ships with an offline database containing 24 columns of data for each ZIP code. You can either keep all 24 variables or filter to just one of these depending on what data you need.
The columns of data provided are: zipcode, zipcode_type, major_city, post_office_city, common_city_list, county, state, lat, lng, timezone, radius_in_miles, area_code_list, population, population_density, land_area_in_sqmi, water_area_in_sqmi, housing_units, occupied_housing_units, median_home_value, median_household_income, bounds_west, bounds_east, bounds_north, bounds_south
Let's begin by using zipcodeR to find all ZIP codes within a given state.
Getting all ZIP codes for a single state is simple, you only need to pass a two-digit abbreviation of a state's name to get a tibble of all ZIP codes in that state. Let's start by finding all of the ZIP codes in New York:
search_state('NY')
## # A tibble: 2,208 × 24
## zipcode zipcode_type major_city post_office_city common_city_list county
## <chr> <chr> <chr> <chr> <blob> <chr>
## 1 00501 Unique Holtsville <NA> <raw 22 B> Suffolk …
## 2 00544 Unique Holtsville <NA> <raw 22 B> Suffolk …
## 3 06390 PO Box Fishers Isl… Fishers Island,… <raw 32 B> Suffolk …
## 4 10001 Standard New York New York, NY <raw 20 B> New York…
## 5 10002 Standard New York New York, NY <raw 34 B> New York…
## 6 10003 Standard New York New York, NY <raw 20 B> New York…
## 7 10004 Standard New York New York, NY <raw 37 B> New York…
## 8 10005 Standard New York New York, NY <raw 35 B> New York…
## 9 10006 Standard New York New York, NY <raw 31 B> New York…
## 10 10007 Standard New York New York, NY <raw 20 B> New York…
## # … with 2,198 more rows, and 18 more variables: state <chr>, lat <dbl>,
## # lng <dbl>, timezone <chr>, radius_in_miles <dbl>, area_code_list <blob>,
## # population <int>, population_density <dbl>, land_area_in_sqmi <dbl>,
## # water_area_in_sqmi <dbl>, housing_units <int>,
## # occupied_housing_units <int>, median_home_value <int>,
## # median_household_income <int>, bounds_west <dbl>, bounds_east <dbl>,
## # bounds_north <dbl>, bounds_south <dbl>
What if you only wanted the actual ZIP codes and no other variables? You can use R's dollar sign operator to select one column at a time from the output of zipcodeR's search functions:
nyzip <- search_state('NY')$zipcode
You can also search for ZIP codes in multiple states at once by passing a vector of state abbreviations to the search_states function like so:
states <- c('NY','NJ','CT')
search_state(states)
## # A tibble: 3,378 × 24
## zipcode zipcode_type major_city post_office_city common_city_list county
## <chr> <chr> <chr> <chr> <blob> <chr>
## 1 06001 Standard Avon Avon, CT <raw 16 B> Hartford…
## 2 06002 Standard Bloomfield Bloomfield, CT <raw 22 B> Hartford…
## 3 06006 Unique Windsor <NA> <raw 19 B> Hartford…
## 4 06010 Standard Bristol Bristol, CT <raw 19 B> Hartford…
## 5 06011 PO Box Bristol <NA> <raw 19 B> Hartford…
## 6 06013 Standard Burlington Burlington, CT <raw 36 B> Hartford…
## 7 06016 Standard Broad Brook Broad Brook, CT <raw 46 B> Hartford…
## 8 06018 Standard Canaan Canaan, CT <raw 18 B> Litchfie…
## 9 06019 Standard Canton Canton, CT <raw 34 B> Hartford…
## 10 06020 Standard Canton Cen… Canton Center, CT <raw 25 B> Hartford…
## # … with 3,368 more rows, and 18 more variables: state <chr>, lat <dbl>,
## # lng <dbl>, timezone <chr>, radius_in_miles <dbl>, area_code_list <blob>,
## # population <int>, population_density <dbl>, land_area_in_sqmi <dbl>,
## # water_area_in_sqmi <dbl>, housing_units <int>,
## # occupied_housing_units <int>, median_home_value <int>,
## # median_household_income <int>, bounds_west <dbl>, bounds_east <dbl>,
## # bounds_north <dbl>, bounds_south <dbl>
This results in a tibble containing all ZIP codes for the states passed to the search_states() function.
It is also possible to search for ZIP codes located in a particular county within a state.
Let's find all of the ZIP codes located within Ocean County, New Jersey:
search_county('Ocean','NJ')
## # A tibble: 32 × 24
## zipcode zipcode_type major_city post_office_city common_city_list county
## <chr> <chr> <chr> <chr> <blob> <chr>
## 1 08005 Standard Barnegat Barnegat, NJ <raw 20 B> Ocean C…
## 2 08006 PO Box Barnegat Li… Barnegat Light, … <raw 33 B> Ocean C…
## 3 08008 Standard Beach Haven Beach Haven, NJ <raw 61 B> Ocean C…
## 4 08050 Standard Manahawkin Manahawkin, NJ <raw 47 B> Ocean C…
## 5 08087 Standard Tuckerton Tuckerton, NJ <raw 51 B> Ocean C…
## 6 08092 Standard West Creek West Creek, NJ <raw 22 B> Ocean C…
## 7 08527 Standard Jackson Jackson, NJ <raw 19 B> Ocean C…
## 8 08533 Standard New Egypt New Egypt, NJ <raw 21 B> Ocean C…
## 9 08701 Standard Lakewood Lakewood, NJ <raw 20 B> Ocean C…
## 10 08721 Standard Bayville Bayville, NJ <raw 20 B> Ocean C…
## # … with 22 more rows, and 18 more variables: state <chr>, lat <dbl>,
## # lng <dbl>, timezone <chr>, radius_in_miles <dbl>, area_code_list <blob>,
## # population <int>, population_density <dbl>, land_area_in_sqmi <dbl>,
## # water_area_in_sqmi <dbl>, housing_units <int>,
## # occupied_housing_units <int>, median_home_value <int>,
## # median_household_income <int>, bounds_west <dbl>, bounds_east <dbl>,
## # bounds_north <dbl>, bounds_south <dbl>
Sometimes working with county names can be messy and there might not be a 100% match between our database and the name. The search_county() function can be configured to use base R's agrep function for these cases via an optional parameter.
One example where this feature is useful comes from the state of Louisiana. Since Louisiana has parishes, their county names don't line up exactly with how other states name their counties.
This example uses approxmiate matching to retrieve all ZIP codes for St. Bernard Parish in Louisiana:
search_county("ST BERNARD","LA", similar = TRUE)$zipcode
## [1] "70032" "70043" "70044" "70075" "70085" "70092"
Try running the above code with the similar parameter set to FALSE or not present and you'll receive an error.
What if you already have a dataset containing ZIP codes and want to find out more about that particular area?
Using the reverse_zipcode() function, we can get up to 24 more columns of data when given a ZIP code.
To explore how zipcodeR can enhance your data & workflow, we will use a public dataset from the National Association of Realtors containing data about housing market trends in the United States.
This dataset, which is updated monthly, contains 10833 observations with current housing market data from the National Association of Realtors hosted on Amazon S3
This is what the data we will be working with looks like:
head(real_estate_data)
## # A tibble: 6 × 40
## month_date_yyyymm postal_code zip_name flag median_listing_… median_listing_…
## <chr> <chr> <chr> <chr> <dbl> <dbl>
## 1 202108 47803 terre haute, in <NA> 219990 0.0732
## 2 202108 48617 clare, mi <NA> 187775 0.0438
## 3 202108 98902 yakima, wa * 246500 -0.0592
## 4 202108 32112 crescent city, fl <NA> 280000 -0.0175
## 5 202108 11764 miller place, ny <NA> 607442. -0.0295
## 6 202108 48317 utica, mi <NA> 257400 0.03
## # … with 34 more variables: median_listing_price_yy <dbl>,
## # active_listing_count <dbl>, active_listing_count_mm <dbl>,
## # active_listing_count_yy <dbl>, median_days_on_market <dbl>,
## # median_days_on_market_mm <dbl>, median_days_on_market_yy <dbl>,
## # new_listing_count <dbl>, new_listing_count_mm <dbl>,
## # new_listing_count_yy <dbl>, price_increased_count <dbl>,
## # price_increased_count_mm <dbl>, price_increased_count_yy <dbl>, …
Note: The data used in this vignette was filtered to only include valid 5-digit ZIP codes as zipcodeR does not yet have a function for normalizing ZIP codes. The full Realtor dataset will have a different number of rows.
We'll focus on the first row for now, which represents the town of Terre Haute, In.
real_estate_data[1,]
## # A tibble: 1 × 40
## month_date_yyyymm postal_code zip_name flag median_listing_… median_listing_…
## <chr> <chr> <chr> <chr> <dbl> <dbl>
## 1 202108 47803 terre haute, in <NA> 219990 0.0732
## # … with 34 more variables: median_listing_price_yy <dbl>,
## # active_listing_count <dbl>, active_listing_count_mm <dbl>,
## # active_listing_count_yy <dbl>, median_days_on_market <dbl>,
## # median_days_on_market_mm <dbl>, median_days_on_market_yy <dbl>,
## # new_listing_count <dbl>, new_listing_count_mm <dbl>,
## # new_listing_count_yy <dbl>, price_increased_count <dbl>,
## # price_increased_count_mm <dbl>, price_increased_count_yy <dbl>, …
The Realtor dataset contains a column named postal_code containing the ZIP code that identifies the town. We'll use this to find out more about Terre Haute than what is provided in the housing market data.
So far we've covered the functions provided by zipcodeR for searching ZIP codes across multiple geographies. The package also provides a function for going in reverse, when given a 5-digit ZIP code. Introducing reverse_zipcode():
# Get the ZIP code of the first row of data
zip_code <- real_estate_data[1,]$postal_code
# Pass the ZIP code to the reverse_zipcode() function
reverse_zipcode(zip_code)
## # A tibble: 1 × 24
## zipcode zipcode_type major_city post_office_city common_city_list county state
## <chr> <chr> <chr> <chr> <blob> <chr> <chr>
## 1 47803 Standard Terre Hau… Terre Haute, IN <raw 23 B> Vigo … IN
## # … with 17 more variables: lat <dbl>, lng <dbl>, timezone <chr>,
## # radius_in_miles <dbl>, area_code_list <blob>, population <int>,
## # population_density <dbl>, land_area_in_sqmi <dbl>,
## # water_area_in_sqmi <dbl>, housing_units <int>,
## # occupied_housing_units <int>, median_home_value <int>,
## # median_household_income <int>, bounds_west <dbl>, bounds_east <dbl>,
## # bounds_north <dbl>, bounds_south <dbl>
You may also be interested in relating data at the ZIP code level to Census data. zipcodeR currently provides a function for getting all Census tracts when provided with a 5-digit ZIP code.
Let's find out how many Census tracts are in the ZIP code from the previous example.
get_tracts(zip_code)
## # A tibble: 11 × 3
## ZCTA5 TRACT GEOID
## <chr> <chr> <dbl>
## 1 47803 000400 18167000400
## 2 47803 000700 18167000700
## 3 47803 001300 18167001300
## 4 47803 001400 18167001400
## 5 47803 001500 18167001500
## 6 47803 001600 18167001600
## 7 47803 001700 18167001700
## 8 47803 010100 18167010100
## 9 47803 010600 18167010600
## 10 47803 010701 18167010701
## 11 47803 010702 18167010702
Now that you have all of the tracts for this ZIP code, it would be very easy to join this with other Census data, such as that which is available from the American Community Survey and other sources.
But ZIP codes alone are not terribly useful for social science research since they are only meant to represent USPS service areas. The Census Bureau has established ZIP code tabulation areas (ZCTAs) that provide a representation of ZIP codes and can be used for joining with Census data. But not every ZIP code is also a ZCTA.
zipcodeR provides a function for testing if a given ZIP code is also a ZIP code tabulation area. When provided with a vector of 5-digit ZIP codes the function will return TRUE or FALSE based upon whether the ZIP code is also a ZCTA.
is_zcta(zip_code)
## [1] TRUE