/1940-singh-census

Every single "Singh" in the 1940 U.S. Census

The 1940 Singh Census

By 1940, there were over a thousand Sikh immigrants, their spouses, and their children living in the United States. Most of the families were mixed-race, with many part of the Punjabi-Mexican community. The majority of these immigrant Sikh men had the last name Singh.

This dataset lists all 962 people with the last name "Singh" in the 1940 U.S. Census, covering a wide cross-section of Sikh Americans.

It's made available by Anirvan Chatterjee from the Berkeley South Asian Radical History Walking Tour.

Datasets

There are two versions of this data, both encoded as standard JSON:

  • 1940_singh_census_basic.json is entirely in the public domain
  • 1940_singh_census_with_geodata.json contains all the public domain data, plus geodata from OpenStreetMap, licensed under the ODbL

Example record

{
   "age" : "19",
   "birthPlace" : "Oklahoma, United States",
   "birthPlaceNormalized" : "Oklahoma, United States",
   "birthYearApprox" : 1921,
   "censusLocation" : "Sup D 3, Maricopa County, Arizona, United States",
   "censusLocationNormalized" : "Maricopa County, Arizona, United States",
   "children" : [
      "Karlena",
      "Jerry"
   ],
   "gender" : "Female",
   "identifier" : "urn:uuid:5B2CC64F-0DE6-39F4-845A-3640B23D7BEA",
   "inferredGroups" : [
      "non-South Asian wife"
   ],
   "maritalStatus" : "Married",
   "name" : "Mildred Singh",
   "raceRecorded" : "White",
   "raceRecordedNormalized" : "White",
   "relationshipToHead" : "Wife",
   "spouse" : "Rala Singh"
},

Data dictionary

Main fields

These fields are in the public domain. All of them come straight out of the U.S. Census, except those specified where I did additional cleanup and tagging work.

  • name: name
  • gender: either "Male" or "Female"
  • maritalStatus: either "Single", "Married", "Widowed", or "Divorced"
  • spouse: name of spouse (in some cases, that spouse will also be in the dataset; look for a person with that name in the same censusLocation)
  • relationshipToHead: relationship to the census head of household, e.g. "boarder," "son" (there are 44 different statuses!)
  • parents: names of parents, father's name typically preceding mother's (in some cases, the parents will also be in the dataset; look for a person with that name in the same censusLocation)
  • siblings: names of siblings
  • raceRecorded: how the census taker recorded their race, often includes abbreviations, e.g. "White", "Black", "Hin", "H"
  • raceRecordedNormalized: my normalized version of the race (e.g. I turn "H", "Hi", and "Hin" into "Hindoo")
  • age: age as of the date of the 1940 census, either an integer string like "21", or "<1" for children under 1 year of age
  • birthYearApprox: approximate birth year (calculated as the age subtracted from 1940)
  • birthPlace: the birthplace listed in the census, most often a country or US state, e.g. "India", "California, United States", "Sri Lanka", "Asia", "West Indian" (sic), "Inida" (sic)
  • birthPlaceNormalized: my attempt at normalizing the birth place to a modern geocodable location
  • censusLocation: the person's location as recorded in the census, often including a district name or number
  • censusLocationCleaned: my attempt at normalizing the census location to a modern geocodable location
  • identifier: a unique identifier for this record, expressed as a UUID URN
  • inferredGroups: one or more of my potentially-incorrect attempts at categorizing people (some of the current groups are "Mexican wife", "married South Asian man", "married South Asian man with kids", and "US-born child")

OpenStreetMap geographic data

The 1940_singh_census_with_geodata.json file also contains extended geodata for the named locations, retrieved via OpenStreetMap's Nominatim. The contents of these fields are © OpenStreetMap contributors, and made available under the Open Database License.

  • birthPlaceGeocodedData: OpenStreetMap data for birthPlaceCleaned
  • censusLocationGeocodedData: OpenStreetMap data for censusLocationCleaned

Almost every location in the dataset contains this geocoded data, except for two cases where a person's birthplace is "West Indies," a location OpenStreetMap can't seem to resolve. (Maybe I should have resolved that to "Haiti" and just called it a day?)

Citation

While the main dataset is in the public domain, you're encouraged to cite the source. For example:

Chatterjee, Anirvan. (2017). 1940 Singh Census. Retrieved from https://github.com/anirvan/1940-singh-census.

If you're using the OpenStreetMap geodata, they ask for a credit.