Meet Walter. His full name was probably Walter Scott Kennedy Rutherford, but sometimes he went by Walter Scott Rutherford, and often his middle names were collapsed to S or S.K. 

The picture above shows Walter in 1921, when he was 43. I’m introducing you to Walter, however, because I believe that when he was 21 he was an enumerator for the 1900 Census. On the census sheets he filled out, he recorded his name as Walter S.K. Rutherford — at least, that’s my interpretation of the cursive script below. 

The same census Walter helped conduct, however, includes no legible record of a Walter S.K. Rutherford in the Bay Area. But there were two Walter S. Rutherfords: a 21-year-old medical student living with his family and their cook in a home they owned in Oakland, and a 60-year-old carpenter who lived in a rented home. This essay will assume that the well-off young medical student is the census enumerator, because he would more likely have the skills to secure the job. This guess was reinforced by multiple references to “Walter Scott Kennedy Rutherford” in documents from the nearby University of California.

“Walter Scott Kennedy Rutherford” in a University of California catalogue. 

“Walter S.K. Rutherford” in a Zeta Psi Directory.

My struggle to identify the Walter S.K. Rutherford who helped conduct the 1900 Census introduces both the man himself and the historical question this essay explores: how do human interactions and experiences shape data collection? Every time Walter’s name was recorded, a choice was made about how he would be encoded into history, decisions that determined how difficult it is to connect each instance of his name to Walter himself. These are the same choices Walter himself made as a census enumerator. This essay uses primary sources and spatial analysis to explore the context in which Walter translated the people of San Francisco Enumeration District 33 into rows on a census sheet.

The top of a census sheet for San Francisco Enumeration District 33.

Each census sheet provides a hint of this context in the metadata section at the top, where enumerators specified the district they were working in, their own name, and the date they filled out the sheet. Though not especially revelatory, this information provides details about the context of data collection that are obscured in many contemporary spreadsheets. I used the dates and addresses on Walter’s census sheets to create the map and animated map below, which traces a rough approximation of his path through the neighborhood as he conducted the census. 

The blue dots approximate the addresses Walter visited while enumerating this neighborhood. Please watch the video, which displays the day each address was visited and the order they were recorded in — disregard the hour part of the time display, which is a very rough approximation used for sequencing the points. The data used were manually transcribed from the census sheets and are linked to here.

The animated map can give an impression of the complexities of collecting census data and what the physical experience might have been like, especially if paired with a close examination of the 1905 Sanborn Insurance Map behind it. He moved relatively quickly through the flats on Stevenson and Jamie Streets, for example, but seemed to spend more time returning to Market Street, which contained more businesses with lodgings above them. While Walter mostly moved through buildings one by one, he would occasionally double back to revisit an address. He apparently finished the majority of his data collection on June 9th, but revisited 1015 Market Street on June 13th, a few days before his June 15 deadline. 

Part of the code used to approximate coordinates for historical addresses. All of my code can be found here.

But the map can’t reflect the actual route Walter took through these blocks, and each point is only a rough approximation of the historical address — though with enough time and unpaid student labor, an accurate geocoding of the 1900 addresses could surely be achieved. Furthermore, the specific times are definitely distorted. A slightly improved version could have scaled the amount of time spent at each address by the number of entries recorded there — some buildings contained many people, and presumably would have taken longer to enumerate. The actual amount of time spent in a household, however, is unknowable. Maybe Walter would have been slowed down by a language barrier while interviewing recent immigrants or struggled to win over a couple hostile to a 21-year-old knocking on their door to ask how long they had been married.  

The instructions provided to census enumerators in 1900 fill in some of these qualitative details, if we assume that Walter followed them. He was given a badge that identified him as a census enumerator and advised to “not lose time (and money)” by loitering while he counted, so we might assume that he moved quickly. The hours displayed on the map, while a horrible indication of when Walter was at an address, do roughly reflect the average amount of time he would have spent at each address if he obeyed the Census Bureau’s recommended 10-hour workday (once more, however, remember that addresses varied widely in how many individuals they contained). 

Ultimately, this investigation could serve as an entry point into a larger analysis of the human experiences behind the 1900 Census data and the social factors that shaped them. It raises several questions that could be productively answered with more time and data: How many enumerators live far from their enumeration districts (Walter was from Oakland, not San Francisco)? How many were, like Walter, wealthy and well-educated? Were interactions always as efficient as the Census Bureau hoped, or were hostility or idle small talk common? Answering these questions would illuminate the social factors that shaped the interactions between census enumerators and respondents, which could in turn influence the data collected. Such insights would help us understand the biases built into historical data and how historical research that builds on these sources might account for them.


1900 United States Census, Enumeration District 33, San Francisco, San Francisco County, California, digital image,  

“U.S., Applications for Seaman’s Protection Certificates, 1916-1940,” s.v. “Walter S. Rutherford” (Born 22 Nov. 1877),

1900 United States Census, Enumeration District 369, Oakland City, Alameda County, California, digital image s.v. “Rutherford, Walter S,”  

1900 United States Census, Enumeration District 62, Sausalito Town, Marin County, California, digital image s.v. “Rutherford, Walter S,”  

“U.S., School Catalogs, 1765-1935,” s.v. “Walter Scott Kennedy Rutherford” (Residence Place: Oakland, Cal.),

“U.S., School Catalogs, 1765-1935,” s.v. “Walter S K Rutherford” (Residence Place: Bacon Block, Oakland, Cal.),

N.a. (1900). “Instructions to Enumerators.” Department of the Interior, Census Office. Retrieved from: