So far, the automated workflows download and process data from the LEHD Origin-Destination Employment Statistics (LODES) and the 5-year American Community Survey estimates, including data necessary to produce ‘Petri Dish’ diagrams of place-based employment by industry and occupation. Developed in collaboration with Utile Architecture and Planning.
Currently, to set up the automated process, you modify config.json with the following parameters:
- project – string
Name of the project. Currently, this only appears as the name of the output geopackage (or results folder, if format is ‘shapefile’ or ‘geojson’). - states – array of strings
Array of states of interest using standard two-letter abbreviations. This can be understood as the 'region' of study in the broadest sense, allowing for analysis of, for example, commuter flows over state lines. - placenames – array of objects
Optional. An object containing towns/cities of interest. It/they must be in the states provided as ‘states’. Each object should have the following two properties...- place – string
Name of each town/city of interest. - state – string
Two-letter abreviation of state including indicated place.
- place – string
- crs – integer
Coordinate reference system EPSG code. - census_unit – string
Either ‘tracts’ or ‘block groups.’ - year – integer
Year of interest. Need to identify ranges for each source. I’ve been using 2021. - format – string
Output format. Current possible values are...“gpkg”: Geopackage.“shp”: Shapefile.“geojson": GeoJSON.
If'shp'or'geojson', non-spatial tables are exported as CSVs.
- datasets – array of strings
List of datasets to download and write. Current possible values are..."lodes": Tables derviced from the LEHD origin-destination employment statistics database."occ": Occupation of civilian employed population 16 and over."ind": Industry of civilian employed population 16 and over.
- census_api – string
Optional. Census API key. It’s good practice to access the Census’s API with a credentialing key, though the scripts will run without one. Request one here.
For example, a config.json for a project focused on Salem, MA that also includes adjacent Beverly, MA as a place of interest (and includes census data for all six states that make up the New England region) would look like this...
{
"project": "salem",
"states": ["MA", "ME", "NH", "VT", "CT", "RI"],
"placenames": [
{
"place": "Salem",
"state": "MA"
},
{
"place": "Beverly",
"state": "MA"
}
],
"crs": 2249,
"census_unit": "tracts",
"year": 2021,
"format": "gpkg",
"datasets": ["lodes", "occ", "ind"],
"census_api": "your_api_key"
}Boundaries of a selected census unit in the state of interest: can be either block groups or tracts.
MULTIPOLYGON
- unit_id – string
The unique identifier (AKA the FIPS code, often called the GEOID) - name_long - string
Place state pair used to uniquely identify tract's place. - pl_name – string
Name of the place including the census geography. - selected – boolean
Whether geography lies within the selected place(s).
Boundaries of, in the case of Massachusetts, all municipalities and in other cases, census designated places in the selected state.
MULTIPOLYGON
- name_long - string
Place state pair used to uniquely identify place. - pl_name – string
Name of the place. - state – string
Two-letter abbreviation of state in which geography falls. - selected – boolean
Whether place is selected.
Table including measures derived from the LEHD Origin-Destination Employment Survey (LODES) data at the given census level.
None. 1-to-1 cardinality with census_unit by “unit_id” in both tables.
- unit_id – string
The unique identifier (AKA the FIPS code, often called the GEOID) - work_res_{MUNI_NAME} – integer
(optional, only present if there is a selected placename)
The number of workers who work in the selected municipality who commute from a home that lies within the given census geography. - res_work_{MUNI_NAME} – integer
(optional, only present if there is a selected placename)
The number of workers who live in the selected municipality who commute to a workplace that lies within given census geography. - pct_w_in_town – float (%)
The % of workers who work in the census geography who also live in the town that the census area is in. - pct_w_in_unit – float (%)
The % of workers who work in the census geography who also live in that census geography. - pct_h_in_town – float (%)
The % of workers who live in the census geography who also live in the town that the census area is in. - pct_h_in_unit – float (%)
The % of workers who live in the census geography who also work in that census geography.
Non-aggregated unit-to-unit flows based on the LODES data.
LINESTRING
- h_unit – string
Census geography of work. 1-to-many cardinality with **census_units **by unit_hd = h_unit - h_selected – boolean
Used to select only commutes from a home in the selected place. - w_unit – string
Census geography of home. 1-to-many cardinality with **census_units **by unit_id = w_unit - w_selected – boolean
Used to select only commutes to workplaces in the selected place. - count – integer
The number of workers commuting from h_unit to w_unit.
Place-to-place (so, municipality-to-municipality) flows. This is much simpler to interpret because it's aggregated to the place.
LINESTRING
- pl_n_h – string
Place name of home. 1-to-many cardinality with **places_{state} **by pl_name = pl_n_h - h_selected – boolean
Used to select only commutes from a home in the selected place. - pl_n_w – string
Place name of work. 1-to-many cardinality with **places_{state} **by pl_name = pl_n_h - w_selected – boolean
Used to select only commutes to workplaces in the selected place. - count – integer
The number of workers commuting from pl_n_h to pl_n_w.
These tables break down employment by occupation at various depths, or degrees of granularity based on ACS 5-year estimates of Occupation by Sex for the Civilian Employed Population 16 Years and Over. place indicates that we're looking at census designated places (generally, cities and towns). unit indicates they’re at the census unit level.
The {depth} suffix indicates whether it’s looking at more generalized or more specific categories (i.e., where in the petri dish hierarchy it sits). Higher depth numbers indicate more specific categories, lower depth numbers are more general.
None. occ_unit_{depth} has 1-to-1 cardinality with census_unit by unit_id in both tables.
Reference this table for columns. They are stored as percentages of the total, so each row should sum to 100.
These tables break down employment by industry at various depths, or degrees of granularity based on ACS 5-year estimates of Industry by Sex for the Civilian Employed Population 16 Years and Older. place indicates that we're looking at census designated places (generally, cities and towns). unit indicates they’re at the census unit level.
The {depth} suffix indicates whether it’s looking at more generalized or more specific categories (i.e., where in the petri dish hierarchy it sits). Higher depth numbers indicate more specific categories, lower depth numbers are more general.
None. ind_unit_{depth} has 1-to-1 cardinality with census_unit by unit_id in both tables.
Reference this table for columns. They are stored as percentages of the total, so each row should sum to 100.