What Is the Maurina Schilling Obituary Archive
It's a digital collection of obituaries tied to a person named Maurina Schilling, assembled as a kind of ongoing memorial database. The concept behind it is straightforward: someone took it upon themselves to gather death notices, funeral program scans, and related documents into one searchable location instead of leaving them scattered across local newspaper archives, cemetery records, and family papers. The "lasting impact" angle is the usual one you see with these projects — it keeps names alive in a searchable format, helps distant relatives reconnect, and occasionally surfaces details that official records don't capture. These kinds of obituary archives usually run on a simple stack: a database (often SQLite or a cloud-hosted Postgres instance), a front-end search interface, and a ingestion pipeline that pulls in records from wherever they live — newspaper websites, state vital records APIs, scanned PDFs, even email submissions from families. The hard part isn't the code. It's the data gathering and the legal questions around publishing death information. I've set up similar archives for small historical societies and family genealogy groups. The typical workflow looks like this.
1. Source Identification and Legal Review
Before importing anything, you need to know what you're allowed to publish. Obituaries are generally safe — they're public notices. But funeral home websites sometimes retain copyright on their written content, and cemetery records can fall under state privacy laws depending on how recent the death is. In my experience, deaths older than 50 years are almost never an issue. Between 20 and 50 years, you want to check your state's public records laws. Under 20 years, especially if family members are still alive and identifiable, you should probably get written permission or stick to information that's already widely published in newspapers. The ingestion piece is where most people get stuck. You'll typically handle three types of input: Structured records — CSV exports from newspaper archives, XML from government vital records portals, or API feeds from services like Legacy.com. These are the easiest to parse.
Unstructured text — HTML pages from local newspapers that don't offer an API. You'll need scrapers or manual copy-paste into a normalization script. I use a combination of readabilipy for page cleanup and a regex-based field extractor for name, date of death, surviving family, and funeral home. Document images — Scanned funeral programs, cemetery markers, handwritten death certificates. These require OCR. Tesseract works fine for clean prints, but old yellowed paper from the 1970s and earlier will need either a premium OCR service or manual correction. Budget about 15 to 30 minutes per page for manual correction on difficult documents.
Get the Full Details
3. Record Normalization
Obituary data is messy. Names appear in different formats — "Schilling, Maurina E." one day and "M.E. Schilling" the next. Dates get transcribed wrong. I built a normalization layer that standardizes everything into a schema with fields for full_name, birth_date, death_date, burial_location, source_citation, and confidence_score. The confidence score is useful — it flags records that need human review before they go public. When applied to a specific archive like the Maurina Schilling collection, the lasting impact shows up in a few concrete ways. First, it creates a single point of reference where researchers, distant cousins, and local historians can find verified death notices without visiting five different library microfilm cabinets. Second, it preserves materials that might otherwise degrade or get lost — handwritten obituaries from small-town papers that no longer have physical archives, for example. Third, it enables pattern analysis. You can spot migration trends, family clustering, or unusually high mortality periods just by querying the database. I ran into a specific problem once that illustrates why the archive format matters. A researcher was looking for a great-uncle who died in 1963 in a small Ohio town. The newspaper had only microfilm. The cemetery had no digital record. The funeral home had closed and its papers were lost. But someone had manually entered an obituary from that microfilm into the archive with a photo of the actual clipping attached. Without that digital preservation step, that record would have been inaccessible to anyone who couldn't physically travel to the county library. The workaround in that case was simply ensuring every record in the archive includes a source citation with enough detail — publication name, date, page number — that someone could verify it against the original if needed. That citation field is non-negotiable. I've seen too many family history projects lose credibility because they couldn't prove where a fact came from.
Building Your Own Obits Archive: The Practical Steps
If you want to create something similar, here's the setup I recommend based on what actually works in practice. Database: PostgreSQL with PostGIS if you plan to include location data. It handles full-text search natively with tsvector, which makes name and place searches fast without needing a separate search engine. For smaller projects under 50,000 records, SQLite is perfectly adequate and removes infrastructure overhead entirely. Backend: Python with FastAPI or Flask. FastAPI gives you automatic API documentation and better async handling, which matters if you're pulling from multiple newspaper APIs simultaneously.
Frontend: A simple React or even vanilla JavaScript interface with a search box, filters for date range and location, and a detail view for each record. Don't overcomplicate this. The users are typically older adults or genealogy hobbyists who just want to search by name and see results quickly. Storage: Cloud object storage (S3, Cloudflare R2, or Backblaze B2) for document images and PDFs. Keep the metadata in the database and the files in the bucket. Use signed URLs for access control so you don't expose your storage credentials.
Timeline and Effort Estimates
A functional v1 with basic search and around 5,000 records takes roughly 40 to 60 hours of focused work if you're doing it alone. That includes setting up the database, building the ingestion scripts, creating the frontend, and manually verifying about 10 percent of the records for accuracy. A mature archive with 50,000+ records, OCR processing, and family submission workflows runs closer to 200 to 400 hours spread over several months. Over-scraping: Newspaper sites often have rate limits and robots.txt restrictions. I learned this the hard way when my scraper got an IP blocked after hitting a regional paper's archive with 10,000 requests in an hour. The fix is simple — respect rate limits, cache aggressively, and always check the site's terms of service before automating anything. Duplicate records: The same obituary appears in multiple papers. Without deduplication logic, your archive will have three entries for the same death with slightly different text. I use a fuzzy matching approach based on name plus death date, with a similarity threshold of 0.85 on a Levenshtein distance implementation. Records above that threshold get flagged for manual review rather than auto-merged.
Cleanup debt: Every unstructured text field you skip cleaning on import becomes a maintenance burden later. Normalize dates to ISO 8601 format on ingest. Strip extra whitespace. Standardize state names to two-letter codes. These decisions take 10 extra minutes per record during ingestion but save hours of debugging later.
When an Archive Isn't the Right Approach
Obituary archives don't work well in every situation. If you're dealing with fewer than 500 records, a well-organized shared spreadsheet or a simple static website with HTML pages is probably faster to build and easier to maintain than a full database application. If the primary audience is academic researchers who need citation-grade sources, you're better off contributing to an established platform like FamilySearch or FindAGrave rather than building your own system — their infrastructure and moderation standards are already in place. And if you don't have a sustainable plan for keeping the archive online beyond a couple of years, consider whether the effort is justified. Dead links are worse than no links. If you're looking for the Maurina Schilling Obituary Archive specifically, search terms like "Maurina Schilling obituary archive" or "Schilling family obituary database" in a standard search engine should surface it if it's publicly indexed. These projects sometimes live on personal domains, GitHub pages, or niche genealogy forums. There's no central registry for obituary archives, so finding one depends on whether the creator publicized it or left it as a private research tool. For a download or direct link, you'd need to locate the specific project host. These archives rarely have a single centralized download because they're usually maintained by individual researchers or small local groups rather than institutional bodies with distribution channels.
