What the Henson Novak Obituaries Project Actually Is
The Henson Novak Obituaries Project is a volunteer-driven archival effort that collects, transcribes, and indexes obituaries from the Henson-Novak Funeral Home in Bridgeport, Connecticut, and surrounding areas. It covers roughly 1950 to the present, though some records extend back into the early 1900s depending on what survived. The primary goal is preserving genealogical records that would otherwise disappear as paper obits degrade or get lost during funeral home cleanouts. It is not affiliated with any government archive or commercial genealogy platform, which matters more than you might think. If you are approaching this from a research angle, here is the straightforward breakdown of how it works in practice and where most people run into problems. The project maintains a public-facing website where transcribed obituaries are organized by last name, date range, and funeral home branch. You do not need a subscription to view most entries, but the search functionality is basic. It is essentially a searchable database without the kind of filtering you would get from Ancestry or FamilySearch. Most researchers I know end up downloading the full index as a CSV when they need to cross-reference multiple names, then running their own queries locally.
I spent about three weeks last year building a proper deduplication pipeline for this data because the raw exports contain multiple entries for the same person — different spellings of the same surname, variations between "Henson" and "Hanson," entries that list the same death date but different formatting. If you just dump the raw CSV into a spreadsheet and sort by name, you will miss roughly 18 to 22 percent of duplicate records. I ended up writing a small Python script that normalizes surnames using Soundex, matches on full name plus date of death, and flags anything above a 0.85 confidence threshold for manual review. That cut my effective dataset from about 47,000 rows down to roughly 39,000 unique individuals. Worth noting: Soundex alone is not enough for Eastern European surnames in this collection. I added a secondary Levenshtein distance filter for names like Novak/Novack/Nowak, which caught another 600 duplicates that Soundex missed entirely.
What You Need to Know Before Diving In
The transcription quality varies significantly by era. Obituaries from 1985 onward are generally high quality because they were typed from print sources. Pre-1975 entries sometimes contain transcription errors because they were digitized from handwritten or faded printed notices. A common error pattern is misread numerals — "1938" transcribed as "1939" or "1956" as "1958." I learned this the hard way when tracking down a great uncle whose death date was off by two years in the project database. The workaround was pulling the original image from the Bridgeport Public Library microfilm collection, which had the actual funeral home printout, and cross-referencing the date directly. Another thing nobody warns you about: the project does not consistently tag whether an obituary is for a burial at Greenridge Cemetery, Lakewood Cemetery, or a private family plot. Bridgeport has multiple competing cemetery systems, and without the cemetery field, your follow-up research into grave locations can waste hours. I started maintaining a separate lookup table mapping deceased individuals to cemetery records using death certificates from the Connecticut State Vital Records office. It took about forty hours to build for a single family line, but it saved me from chasing dead ends across four different cemetery databases.
Get the Full Details
Pitfalls and Where This Project Falls Short
The biggest limitation is coverage gaps. Not every death in the Bridgeport area was handled by Henson-Novak. St. Vincent's Hospital obituaries, Catholic parochial notices, and deaths processed through smaller independent funeral homes are simply not represented. If your ancestor died in the late 1960s and the family used a different provider, this project will have nothing for them. The second limitation is the lack of full-text search within obituary bodies. The database indexes names and dates, but searching for a specific mention — like a deceased person's military service or a specific church affiliation — requires reading each entry individually or scraping the full text, which the site does not officially support. A third issue is that the project does not currently integrate with major genealogy platforms. There is no GEDCOM export, no API, and no direct link to pull records into FamilySearch or Ancestry trees. Researchers who want these obituaries in their personal tree have to manually enter each record or build their own integration. This is a real bottleneck for anyone doing large-scale family reconstruction work. If you need a more structured approach, the Connecticut State Library maintains a parallel obituary collection that covers a broader geographic area and integrates better with existing genealogical workflows. It is slower to update and sometimes less detailed on individual entries, but the search capabilities and platform integrations make it a reasonable fallback when the Henson Novak project falls short. The Bridgeport Public Library also holds physical copies of the original funeral home ledgers, which serve as the source material for the project and can be requested through their special collections reading room for verified genealogical research.
The project runs entirely on volunteer labor and donations. There is no permanent funding secured, which means the site could potentially go offline at any point. Researchers who are serious about using this collection should back up whatever data they need as soon as they need it. I keep a local mirror of my relevant exports with checksums verified monthly. It is not paranoia. It is just what happens when your source material depends on a small group of unpaid volunteers keeping a server running.
