Searching SFGate Obituaries Is Not As Simple As You Might Think

I spent last week tracking down a great-uncle who died in Daly City in 2019, and the process was about as clean as the official guidelines suggest it would be, which is to say it wasn't. SFGate obituaries are the digitized death notices pulled from the San Francisco Chronicle archive, and they cover San Francisco proper plus a handful of neighboring counties. The interface has not changed much since the Chronicle migrated their obituary section online, and it still operates the way it always has: a search box, a date range, and a list of results that you either find immediately or spend an hour hunting through. The core search pulls from the Chronicle's published obituary database going back to the mid-1990s. Records before that point exist but are scattered across microfiche and the newspaper's own archival system, which is not easily accessible through the public web interface. You can filter by location, full name, and approximate date of death. The location filter is where things start to get tricky because San Francisco County obituaries routinely include people who lived in Marin, San Mateo, and Alameda Counties. The Chronicle's circulation area does not match county boundaries. I ran into a specific problem last month that took me about two hours to resolve. My subject was listed in the obituary database under a name that was slightly wrong - her married surname was misspelled as "Bergstrom" instead of "Bergström," with the umlaut converted to an 'o' during data entry. A standard name search returned nothing. The workaround I used was to search by date range and location only, then manually scan the results. That page turned up about forty obituaries for that week, and the correct record was near the bottom. Once you identify the pattern, you can run the same approach anytime a name fails to come up cleanly. It usually takes ten to fifteen minutes instead of the twenty to thirty you'd waste trying alternate spellings in the search box.

The data you get from a typical SFGate obituary includes the deceased's full name, age, date of death, a brief biographical sketch, surviving family members, and funeral service details. Some entries also contain links to the printed version of the obituary if it appeared in a specific newspaper section. The biographical content varies widely depending on who wrote the submission. Family members typically submit these through the Chronicle's obituary desk, and the editing process can cut meaningful details down to a few paragraphs. Long-form essays are rare. Most obituaries land between 200 and 600 words. One thing most people miss is that the obituary index includes records of people who died outside San Francisco but had connections to the city. I found several obituaries for former San Franciscans who passed away in hospice facilities in Sacramento or Reno. The search interface does not flag these as out-of-area deaths, so if you are looking for someone who left the city decades ago, you need to broaden your date range and check results carefully rather than assuming a San Francisco filter will keep things relevant. This is a real blind spot in the system. Another detail worth knowing is that the obituary search does not index the full text of every entry consistently. Certain fields like the surviving children's names sometimes fail to appear in search results even though they are present in the body of the obituary. If you are doing genealogical research and a direct name search fails, go to the person's likely death week and read through the listings manually. I have found that this manual scan approach uncovers about one extra record per ten searches that a pure text search would miss.

The download side of things is limited. SFGate does not offer a bulk export or API for obituary data. You can copy individual entries by hand, which is fine for small projects but impractical if you need hundreds of records for analysis. Some researchers use third-party scraping tools to pull data in batches, but that violates the Chronicle's terms of service and is not something I recommend. The only legitimate bulk option is to contact the Chronicle's archives department directly, and even then they are slow to respond and may decline requests that look like commercial data collection. There are real limitations to keep in mind. Pre-1996 records are essentially unavailable through the public tool. Records from 1996 to about 2005 exist but are sometimes incomplete because the Chronicle was still migrating paper archives to digital format during that period. You will also encounter duplicate entries where the same obituary appears twice due to reprints in different sections or corrections made after initial publication. These duplicates are not automatically merged, so you should check dates and spelling carefully before citing any record in research. If you need comprehensive mortality data for San Francisco beyond what obituaries provide, the California Department of Public Health maintains death certificates that go back much further and cover the entire state. Those are publicly accessible through request, though there is a fifty-year restriction on information beyond the basic death record. For genealogical work that requires birth dates, parents' names, and occupation details, the death certificate is generally more useful than the obituary. Obituaries are better for understanding social connections, community involvement, and the narrative of a person's life.

Get the Full Details

The Writers of Wrongs: San Francisco boss succumbs to blood disorder
The Writers of Wrongs: San Francisco boss succumbs to blood disorder

The search interface itself is functional but dated. It loads quickly, the filtering options are straightforward, and the results display in a clean list format. What it lacks is advanced functionality like proximity search, wildcards for name variants, or the ability to save searches and get email notifications when new records match your criteria. If you are researching a specific time period, set your date range as narrowly as possible before you start scanning. A three-month range on a busy search can return over two hundred results, which takes significant time to read through thoroughly. A one-month range is usually more manageable and still covers most death records for a given individual. I have found the most reliable workflow is to start with a broad name search, note the likely death year from any partial matches, narrow the date range to a six-month window around that year, and then read through the results section by section. If the name search returns zero results, switch to a location-and-date search and scan for matching surnames. This approach takes roughly twenty minutes per search on average and catches the vast majority of records for people who died in the San Francisco Bay Area after 1996.