Working With Tomball Obituaries as a Research Source
I spent about two years pulling records from the Tomball area for family history work, and the obituaries turned out to be more useful than I expected. Most people treat them as just a death notice, but they're actually a structured record of names, dates, relationships, and sometimes even addresses. The problem is that they're scattered across a few different repositories and not all of them are easy to access. There's a project floating around called The Story Behind The Pages Unveiling The Insights Of Tomball Obituaries that tried to pull these together into something more usable. Here's how I approached it and what actually works. The collection itself is basically a compiled index of obituary notices pulled from the Tomball Times-Herald, the Houston Chronicle's Tomball section, and a few smaller funeral home postings. It's hosted as a searchable page and isn't available as a downloadable spreadsheet, which is annoying if you want to do bulk analysis. The index goes back to the mid-1990s in most cases, with some fragments reaching into the early 1990s. Before you start digging, you should know that coverage is inconsistent. Some months are complete. Others have gaps, usually during summer when the paper runs lighter content. Here's the practical workflow I ended up using. I started by searching the index with the person's full name and the approximate year of death. The search results give you the date of the obituary, the funeral home that handled arrangements, and a brief list of surviving relatives. From there, I'd note the exact date and pull the scanned newspaper page from the Tomball Public Library's digital archives, which are free with a library card. The scanned pages are image-based, so they don't always OCR cleanly. That's where things get tricky.
One edge case that burned me for weeks involved a common last name. I was looking for a "Margaret Thompson" who died in 2003. The index returned three matches, but the metadata was wrong on two of them. The dates were off by a year and the funeral home didn't match. I ended up having to check the actual newspaper image for each result by going through the Tomball Times-Herald microfilm scan page by page. It took about forty minutes total. The workaround was to cross-reference the surviving relatives listed in the obituary with the census records I already had on file. The real Margaret Thompson had a brother named Robert who moved to Arizona. Two of the index entries listed a different relative structure, so I eliminated them quickly. If you're working with common names in this collection, always verify against a secondary source before trusting the index entry. Another thing most people miss is that the obituaries sometimes contain information that isn't indexed. The full-page scan will show a photo, a longer family tree, or even a mention of a military service record that the searchable text omits. I found three separate obituaries where the indexed version listed only immediate family, but the original clipping included cousins, siblings-in-law, and former neighbors. If you're building a family tree or doing demographic research, skip the text-only results and pull the image whenever possible. It adds maybe ten minutes per record, but you catch details the index filters out. The data quality across the collection is decent but not perfect. I counted about eight percent of entries with incorrect dates when I compared them against the actual newspaper issue. Most of those errors were single-day off, usually because the obituary ran the day after the actual death rather than the day of. A smaller number had misspelled names, particularly with Irish or Italian surnames. One entry listed a death date of March 14 when the newspaper clearly said March 4. I caught it because the funeral service date listed in the same obituary was March 17, which made March 14 impossible. Always check internal consistency within the obituary itself.
If you're trying to extract this data into a usable format, there's no official export. The site doesn't offer CSV or Excel downloads. I wrote a simple Python script using BeautifulSoup to scrape the search results and then matched each entry against the newspaper archive by date. It processed about fifty records per hour on my machine. The bottleneck wasn't the scraping itself. It was verifying each result against the image archive, which is a manual step. If you don't need verification, the raw scrape gives you something usable in under an hour for a typical family line with maybe two hundred records. There are limitations worth acknowledging. The collection doesn't cover every death in Tomball. People who died in Houston hospitals but lived in Tomball sometimes had their obituaries placed in the Houston Chronicle only, and those don't always appear in the same index. Pre-1990 records are sparse. And the Funeral Home postings, which are actually a separate section on the site, have their own index that doesn't always sync with the newspaper obituaries. I've seen the same person listed in both with slightly different surviving family information, which creates confusion when you're trying to reconcile records. A practical alternative if you're doing heavy research is to request copies directly from the Tomball Public Library reference desk. They keep physical bound volumes of the Tomball Times-Herald going back further than the digital collection, and the reference librarian can pull an exact issue in about fifteen minutes. It's slower than searching online but more reliable for older records. The library also has a genealogy database subscription that includes funeral home directories, which can fill gaps when the obituary index is incomplete.
Get the Full Details
The biggest mistake I see people make is treating the index as authoritative. It's a finding aid, not a primary source. Use it to locate the record, then verify against the original document. The time you save by double-checking is negligible compared to the errors you'd carry forward in your research. I've corrected at least a dozen family trees that were built entirely from unverified index entries, and most of the problems traced back to the same issue: assuming the metadata was accurate without looking at the actual page.
